NYSBA Members Propose Regulatory Fix to Address Multiple Guarantors

I am pleased to submit the attached report of the New York State Bar Association Tax Section. The report makes proposals regarding the treatment under section 956(d) of the Internal Revenue Code of debt obligations of a U.S. person that are supported by a guarantee or pledge of assets of multiple controlled foreign corporations ("CFCs").

Section 956(d) can be read to require each CFC that provides a guarantee or asset pledge in support of an obligation of a U.S. person to be treated as holding the entire obligation for purposes of section 956. Therefore, there is a risk under the current rules that a United States shareholder of multiple CFCs could be required to recognize an inclusion under section 956 in an amount greater (in some cases, many times greater) than the amount of the debt obligation supported by the guarantee or asset pledge. This so-called "multiple guarantor problem" thus can give rise to taxable income far in excess of any conceivable economic benefit realized by the taxpayer.

The report recommends the promulgation of regulations under section 956 that would to address situations where multiple CFCs guarantee a single debt obligation. The report refers to these debt obligations as "section 956(d) group obligations." We recommend that the regulations provide that the total amount taken into account on account of a section 956(d) group obligation under section 956 by a United States with respect to all of its CFCs would be no more than the amount of the obligation.

While there are several potential approaches these regulations could take, we believe the best approach is to cap the amount determined under section 956 for a United States shareholder with respect to any one CFC at:

At the end of the report, we have provided draft regulatory language to illustrate how such a rule might work in detail. As will become apparent from the report, this is truly an area where the "devil is in the details," and the draft language represents our attempt to help flesh out those details in a concrete way. We appreciate your consideration of our recommendations. If you have any questions or comments regarding this report, please feel free to contact us and we will be glad to discuss or assist in any way.

This report of the Tax Section of the New York State Bar Association¹ recommends a regulatory rule (along with illustrative draft language) that is intended to resolve a longstanding technical issue under section 956 of the Internal Revenue Code of 1986, as amended (the "Code").² Specifically, in the preamble to proposed Treasury regulations under section 956 dated September 2, 2015 (the "Preamble"), the Treasury Department ("Treasury") and the Internal Revenue Service (the "Service") indicated that they are considering the adoption of regulations to address the case where a single debt obligation of a United States person benefits from credit support (either in the form of a pledge of assets or stock of a CFC or a guarantee, each of which will be referred to as a "guarantee" for purposes of this report) of more than one "controlled foreign corporation" ("CFC").³ Currently, neither section 956 nor the regulations thereunder address this fact pattern directly. The statutory language of section 956(d), however, may be read to require a United States shareholder of such CFC to include in income an amount that could be many times larger than the amount of the debt obligation in question -- and therefore many times larger than any possible economic benefit actually realized by the United States shareholder.

We believe that such an outcome is inconsistent with economic reality and with the overarching policy of section 956, which is to capture value that has been repatriated without a formal dividend declaration from a CFC. Well advised taxpayers almost always are able to avoid non-economic inclusions under section 956(d), when they choose to do so, through the use of straightforward and well established "fixes" that are established in regulations and have generally been accepted by the financial markets. For less well-advised taxpayers, however, the possibility of section 956 inclusions in excess of the actual debt amount serves as a trap for the unwary that can give rise to tax liabilities far in excess of the economic benefit that a U.S. person may have derived from a CFC's guarantee. In addition, as with any non-economic outcome, the creation of multiple inclusions under section 956 attributable to a single debt obligation may be manipulated by taxpayers as a tool in planning -- e.g., to effect the repatriation of foreign tax credits without the corresponding repatriation of cash, or to offset net operating losses that otherwise might expire and replace them with "previously taxed income" ("PTI") balances. We therefore applaud the interest that Treasury and the Service have shown in addressing this issue.

Part II of this report first provides a general background and overview of the workings of section 956, as well as a discussion of the concern that section 956 is intended to address and how section 956 interacts with other Code sections. Part III of this report discusses issues raised in the Preamble and other issues that we considered when developing the draft regulatory language suggested by this report. Part IV of this report sets out the framework of the analysis behind our recommendation. Part V of this report contains a more detailed technical discussion of the mechanical operation of our recommendation, and the Appendix contains a draft of the suggested regulatory language, including examples.

II. SUMMARY RECOMMENDATIONS.

We recommend the promulgation of regulations under section 956 that would to address situations where multiple CFCs guarantee a single debt obligation. We refer to these debt obligations as "section 956(d) group obligations." We recommend that the regulations provide that the total amount taken into account on account of a section 956(d) group obligation under section 956 by a United States with respect to all of its CFCs would be no more than the amount of the obligation.

While there are several potential approaches these regulations could take, we believe the best approach is to cap the amount determined under section 956 for a United States shareholder with respect to any one CFC at:

III. SECTION 956 AND SUBPART F GENERALLY.

Subpart F income inclusions are limited in two important respects: First, under section 952(c)(1), the subpart F income of a corporation during any taxable year cannot exceed the corporation's current earnings and profits (referred to herein as current "E&P") for such taxable year. In other words, the "deemed dividend" of a subpart F income inclusion can be "paid" only out of current E&P. This limitation is more taxpayer-favorable than a pure deemed dividend construct necessarily would require, since actual cash dividends may be paid out of current or accumulated E&P. Second, once there has been a subpart F income inclusion, section 959 provides generally that there will be no further income inclusion later when amounts allocable to that E&P either actually are distributed or, as discussed below, are deemed distributed by virtue of section 956. This result is achieved by recording subpart F income inclusions in an account for PTI, the balance of which account is then "paid down" by later nontaxable distributions or deemed distributions. PTI balances are specific to individual shares of stock, so that a purchaser of stock succeeds to the portion of the seller's PTI balance allocable to that stock. Importantly, PTI is treated as distributed first -- i.e., in priority to other E&P.

In contrast to subpart F income inclusions, section 956 inclusions may be deemed paid out of both current and accumulated E&P, and are not limited to any specific type of income realized by a CFC. This reflects the fact the section 956 is intended to capture amounts that economically are similar to actual repatriations of cash or property, as opposed to capturing only subpart F income, where the concern is that the income has not been repatriated. As a technical matter, section 956 applies to investments (or deemed investments) in "United States property," which term includes certain tangible property located in the United States, stock or obligations of a United States shareholder of the CFC or of a U.S. corporation in which such United States shareholder owns 25 percent or more of the voting power, and rights to use certain types of intangible property acquired or developed by a CFC for use in the United States.⁷

The theory underlying section 956 is that these types of investments in United States property put the assets of the CFC at the disposal of the United States shareholders in a manner that is economically similar to a dividend.⁸ If a CFC loans money to a U.S. parent, or invests in U.S. parent stock, for example, then the U.S. parent has availed itself of the CFC's assets and should be taxed in the same manner as if those assets had been distributed to the U.S. parent. Similarly, if a CFC invests in tangible property located in the United States, there is an implicit assumption that such investment is unlikely to be for use in the CFC's (presumably non-U.S.) trade or business, and is more likely to benefit a United States shareholder that might have made that investment with the proceeds of an actual cash dividend.

The specific amount of a section 956 inclusion is determined under section 956(a) by reference, in the first instance, to the relevant United States shareholder's pro rata share of the lesser of :

As in the case of a subpart F income inclusion, a section 956 inclusion during any taxable year gives rise to a PTI balance equal to the amount of such inclusion. The PTI balances for subpart F income inclusions and for section 956 inclusions are maintained separately, and the two different balances are referred to herein as "subpart F income PTI balances" and "section 956 PTI balances." These two balances are calculated on a shareholder-by-shareholder basis and coordinated in the following manner: First, subpart F income PTI balances for any given taxable year are established to the extent that there are subpart F income inclusions occurring for that taxable year.¹⁰ Then, any section 956(a) amounts for that same year (i.e., investments in United States property in excess of section 956 PTI balances) are deemed paid out of subpart F income PTI balances to the extent such balances exist for the current year or prior years and have not been "paid down" previously, ¹¹ with the result that amounts may be transferred from a subpart F income PTI balance to a section 956 PTI balance.¹² (That transfer is not a taxable event for United States shareholders.) Section 956(a) amounts in excess of available subpart F income PTI balances give rise to a taxable section 956 inclusion for United States shareholders,¹³ and give rise to a corresponding increase in the section 956 PTI balance. Finally, any cash distributions made during a taxable year are deemed paid first out of any available section 956 PTI balances,¹⁴ and then out of any available subpart F income PTI balances.¹⁵ To the extent of those PTI balances, the repatriation of cash from a CFC does not constitute a taxable dividend.¹⁶

The coordination of PTI balances under subpart F is illustrated in the following examples: Consider a CFC wholly owned by a single U.S. parent corporation. At the beginning of Year 1, the CFC has no investments in United States property, no subpart F income PTI balances and no section 956 PTI balances. The CFC has sufficient current and accumulated E&P such that E&P will not be limiting factor to the inclusion of subpart F inclusions and section 956 inclusions.

In Year 1, the U.S. parent recognizes a subpart F income inclusion of $2000 from the CFC, which gives rise to a subpart F income PTI balance of $2000. Also in Year 1, the CFC loans $2500 to the U.S. parent, which constitutes an investment in United States property giving rise to a section 956(a) amount of $2500. The $2500 section 956(a) amount is allocated first against the $2000 subpart F income PTI balance, with the result that the subpart F income PTI balance is reduced to zero, and the section 956 PTI balance is increased to $2000. That shift of income from one PTI balance to the other gives rise to no additional taxable income for the U.S. parent. The remaining $500 of the section 956 amount, however, does give rise to $500 additional taxable income to the U.S. parent in the form of a section 956 inclusion, and increases the section 956 PTI balance to $2500.

Then, in Year 2, there is no change to the CFC's investments in United States property, but there is an additional subpart F income inclusion of $700. In that year, the CFC also makes a cash distribution to the U.S. parent of $5000. The treatment of these transactions is that, first, the U.S. parent's subpart F income PTI balance is increased to $700, to reflect the inclusion of $700 of taxable income by the U.S. parent. Then, the $5000 cash distribution is allocated first against the pre-existing $2500 section 956 PTI balance from Year 1, and then against the $700 subpart F income PTI balance from the current Year 2. Accordingly, $3200 of the distribution gives rise to no additional taxable income to the U.S. parent. The remaining $1800 is a taxable dividend. In addition, because the CFC's investment in United States property remains at $2500, at a time when the CFC's section 956 PTI balance has been reduced zero, the CFC's section 956(a) amount for Year 2 is again $2500. Because none of that section 956(a) amount is offset by a subpart F income PTI balance following the $5000 distribution, the U.S. parent has an additional taxable section 956 inclusion in the amount of $2500, and the section 956 PTI balance is restored to $2500. The U.S. parent thus has included a total amount of taxable income for Years 1 and 2 of $7500, which equals the sum of the total cash paid or loaned to the U.S. parent. If the U.S. parent were then to sell half of its CFC stock to a third party, that third party would take the stock with a section 956 PTI balance of $1250.

Prior to 1980, regulations under section 956(d) simply restated the rule of the statute, and explained that the amount taken into account under section 956(d) is the unpaid principal amount of the U.S. person's debt that is the subject of the guarantee.¹⁸ The regulations at that time were silent on the question of so-called indirect pledges, which occur in cases where a United States shareholder avails itself of a CFC's balance sheet, not by having the CFC pledge assets or make a guarantee, but instead by pledging the stock of the CFC to a lender, usually with various negative covenants to ensure that assets are not stripped from the CFC and that the CFC stock retains its value as collateral. The Service, however, did address the question of indirect pledges in Revenue Ruling 76-125¹⁹ and concluded that a pledge of CFC shares coupled with negative covenants properly is treated as a guarantee for purposes of section 956. The revenue ruling addresses the case of the pledge of stock of a single CFC, and merely says that the section 956(a) amount is equal to the amount of the CFC's E&P "available for a distribution."

In Ludwig v. Commissioner,²⁰ the Tax Court rejected the Service's position as expressed in Revenue Ruling 76-125, which caused the Service to amend the regulation under section 956(d) in 1980. The taxpayer in Ludwig was the 100-percent shareholder of a foreign corporation, Oceanic, that had $5 million in undistributed E&P. The taxpayer borrowed $100 million from third party banks and used its stock in Oceanic as collateral for the borrowing. In addition, the banks imposed certain negative covenants on the taxpayer, in order to prevent the taxpayer from compromising the value of his interest in Oceanic. The Service argued that the taxpayer should recognize a section 956 inclusion equal to Oceanic's undistributed E&P because the taxpayer's pledge of Oceanic stock was equivalent to Oceanic being a guarantor in respect of the borrowing (the pledge of stock in and of itself did not fall within the then-existing language of section 956 and the regulations thereunder, which would have captured only the pledge of a CFC's assets). The Tax Court disagreed. Although the Tax Court acknowledged that the taxpayer had realized a benefit from the Oceanic stock, the court held that "[n]either a pledge of the[CFC's] stock to secure a shareholder's loan nor the listing of such stock on a balance sheet as evidence to support a loan, constitutes an investment of earnings in United States property."

In response to the Ludwig decision, the Service issued Treasury regulation section 1.956-2(c)(2) in 1980, which regulation provides that:

Other than in the case of a taxpayer that has made an unfortunate misstep, we doubt that the rules of section 956(d) have generated much net revenue for the fisc since Treasury regulation section 1.956-2(c)(2) was issued in 1980. This view is due to the references in the regulations to a pledge of "at least 66 2/3 percent of the total combined voting power of all classes of stock entitled to vote" and "one or more negative covenants or similar restrictions." Those provisions generally are interpreted as creating a safe harbor. In our experience, therefore, it is uncommon for U.S. parents -- knowingly and in the absence of a larger tax planning objective -- to require their CFC subsidiaries to guarantee parent borrowings directly or to pledge assets in support of those borrowings. Instead, it is routine for a U.S. parent to pledge 65 percent of a CFC's voting stock (or some substantial amount of voting stock less than 66 2/3 percent) and 100 percent of any non-voting stock of the CFC, provide the lenders with negative covenants, and take the well-established view that section 956 does not give rise to a section 956 inclusion in such circumstances. It is our collective experience that this practice is widely understood and accepted among lenders, and that lenders do not generally require any additional credit support from CFCs that might give rise to a section 956 inclusion.

For these reasons, section 956 inclusions arising from direct and indirect guarantees and pledges by a CFC constitute primarily a trap for the unwary, and the issues discussed in this report concern taxpayers that in the vast majority of cases have fallen inadvertently into a tax problem that could have been avoided with appropriate tax planning. Again, well advised U.S. parent borrowers generally do not allow their CFCs to provide guarantees and pledge assets in support of U.S. parent borrower debts. They typically pledge less than 66 2/3 percent of the voting stock of the CFCs instead and avoid section 956 inclusions altogether. To the extent that a taxpayer has knowingly created a taxable dividend inclusion through the operation of section 956(d), that inclusion most likely would be part of a larger tax planning strategy.

This reading is based on the literal language of section 956(d), which provides that:

The problem with this reading of section 956(d) is illustrated in the following examples.

Example (1). A U.S. person (A) borrows $100 from unrelated third party lenders on January 1. The borrowing is outstanding at all points during the taxable year, and there is no increase or decrease in the principal amount of the borrowing during that time. A's borrowing is guaranteed by its wholly owned foreign subsidiary, B, which is a CFC. B has E&P of $300, and a subpart F income PTI balance of $30.

Example (2). Assume the facts of Example (1), except that A's borrowing is guaranteed by three wholly owned foreign subsidiaries that are CFCs (C, D and E). Each has a subpart F income PTI balance of $10 and E&P of $100.

In Example (1), A's section 956 inclusion equals $70, calculated as follows: The principal amount of the obligation in respect of which B serves as guarantor is $100. B has a subpart F income PTI balance of $30, and the section 956 inclusion is the excess of the outstanding debt over PTI, or $70. This result comports with the policies underlying section 956, because it measures properly the potential economic benefit realized by A that is attributable to B's credit support. If A were to default on the obligation and B were called upon to pay $100 to satisfy A's debt, the economic benefit to A would be the same as if A had received a $100 distribution and used the distribution proceeds to pay the debt. In that case, only $70 of the distribution would have been taxable income to A (because of the $30 subpart F income PTI balance), and that is the same result achieved through the section 956 inclusion.

In Example (2), A's taxable income from multiple section 956(a) amounts would arguably be $270. If each of C, D and E are treated as having effected separate, standalone purchases of the entire $100 debt, then each purchase gives rise to a section 956(a) amount, and in each case the section 956 inclusion is the excess of the amount of the debt obligation over the $10 of subpart F income PTI -- for a total income inclusion of $270.

Of course, the result in Example (2) fails to comport with economic reality. If C, D and E ever were called upon to pay on their guarantees, their payments would be coordinated so that the lender received only the $100 amount of the loan and not $270 or $300. The economic effect to A of a satisfaction of the guarantees would be the same as in Example (1) -- that is, A is in the same economic position as if it had received only $100 in distributions.

In our view, the example above illustrates the need for a rule coordinating guarantees of multiple CFCs so that they are judged by their aggregate effect on the United States shareholder and are not viewed for tax purposes as separate unconnected events. The above-quoted statutory language from section 956(d) expressly contemplates that Treasury and the Service will use their regulatory power to effect a rational system under that Code section, and we believe that a proper coordination rule thus is a part of that regulatory mandate. Of course, the only recent regulatory action in respect of section 956(d) was the change to the regulations in response to the Ludwig case, discussed above. As described in the Preamble, Treasury and the Service are considering the possible adoption of a coordination rule under section 956(d). We applaud that development, but also recognize that the task of formulating an appropriate rule brings with it a fair amount of complexity. In Part V, we suggest a rule (and related draft regulatory language in the Appendix as an example of how such a rule might work in detail) for the Service's consideration, but first, Part IV discusses some of the issues we considered when formulating the draft language and how we attempted to resolve them.

IV. FRAMEWORK.

In considering how best to achieve an appropriate coordination rule (and how best to balance potentially competing concerns such as administrability and precision), it is worth keeping in mind that most taxpayers who wish to do so will continue to be able to avoid section 956 inclusions with respect to CFC credit support altogether by opting to pledge less than 66 2/3 percent of the CFCs' voting stock. This reality significantly mitigates the concern that taxpayers might take inappropriate advantage of some technical feature of a coordination regime, say, to reduce a section 956 inclusion that could have been avoided in its entirety through appropriately limited share pledges. Similarly, to the extent taxpayers seek to use multiple guarantors as a tax-planning tool to create artificial inclusions, a coordination rule could go a long way towards preventing the creation of noneconomic tax benefits.

Consistent with the discussion in the Preamble, we considered several alternative means of achieving coordination though allocating a single debt obligation among its several different CFC guarantors at the time the obligation is entered into. It seemed to us that any coordination rule capable of operating properly within the larger framework of subpart F would have to reflect the effects of a section 956 inclusion at the level of specific individual CFCs, because the resulting changes to PTI balances would need to be reflected as changes to attributes of specific CFCs in order to be taken account of properly in future years. Accordingly, a rule for allocating debt among individual CFCs, or something very similar, seemed to us to be necessary.

On the other hand, we believe that any debt allocation rule will be artificial to some degree, in that any such rule is unlikely to reflect the legal reality arising when there are multiple guarantors of the same debt obligation. In the most common case where there are multiple guarantors of a single debt obligation, each guarantor potentially may be called upon to pay the entire amount of the obligation, and thus each has a contingent liability for that entire amount. At least in commercial third-party financings, it is quite uncommon for guarantors to divide a debt liability among themselves so that each guarantor is liable to a lender only for a portion of the debt.²³ If guarantors did so, however, section 956(d) as currently drafted might not give rise to multiple inclusions. Accordingly, it is necessary to choose an allocation mechanism without the benefit of being able to look to some division of the debt based on the legal arrangements among the borrower, the guarantors and the creditor. The question then becomes one of choosing the appropriate criteria by which to allocate a debt obligation.

We considered allocating debt among CFCs by reference to the CFCs' relative fair market values. Although that approach has some intuitive appeal (because it is based somewhat on the economic value provided by the different guarantees), we rejected the approach on two grounds. First, most CFCs are privately held companies with illiquid stock, and we were concerned that it might be difficult to determine relative fair market values of CFCs in an accurate and easily verifiable manner. The problem would be exacerbated in the context of a United States shareholder that has accidentally triggered a section 956(d) inclusion and does not become aware of the fact until several years later, at which point the United States shareholder presumably would be required to determine values on a retrospective basis. Second, this approach is not consistent with the general mechanics of section 956, which do not look to the value of the guarantee as a factor relevant to the calculation of a section 956 amount.

We also considered allocating debt obligations evenly among CFCs (so that, if there are three CFC guarantors, for example, each is allocated a third of the debt obligation), but that approach struck us as being both arbitrary and subject to manipulation. It would be simple for taxpayers to "dilute" the section 956 inclusion from a single CFC guarantor simply by adding other guarantors, potentially with minimal or no positive E&P balances (a concern noted in the Preamble).

Ultimately, we concluded that the real problem with each of the above approaches is that they are not based on factors that actually drive the calculation of section 956 amounts. A simple example illustrates the potential pitfalls that may arise under an allocation regime that is not based on such factors: Assume that a United States Shareholder borrows $1000 utilizing a guarantee from each of its three wholly owned CFCs. The CFC guarantors have no other investments in United States property and guarantee no other debt. The CFC guarantors have the following tax profiles: (i) the first CFC guarantor has a positive subpart F income PTI balance of $500 that accounts for all of its current and accumulated E&P; (ii) the second has no positive PTI balance and a large E&P deficit; and (iii) the third has an E&P balance of $2000 and no positive PTI balance. An allocation of one third of the obligation, for example, to each CFC (e.g., based on dividing the debt evenly, or as a result of the fact the CFCs are of equivalent value) would give rise to: (i) a $333.33 reduction in the subpart F income PTI balance for the first CFC with no taxable section 956 inclusion; (ii) no section 956 inclusion from the second CFC because of the E&P deficit; and (iii) a taxable section 956 inclusion of $333.33 from the third CFC.

Each of these three results strikes us as arbitrary because the effects of the allocation of debt to a CFC in the above example are independent of the criteria by which the allocations are made. This disconnect gives rise to a level of unpredictability in the operation of section 956(d) that we find difficult to justify. As an initial matter, we believe that the United States shareholder should have been required to take the entire $1000 borrowing into account as a section 956(a) amount, since the members of the CFC group have in the aggregate sufficient "applicable earnings" (as defined in section 956(b)(1)) to support a section 956(a) amount of $1000 (and we see no reason to offset one CFC's E&P deficits against another's positive E&P balance). Accordingly, a better approach would be to allocate debt among CFC guarantors in accordance with their positive E&P balances -- such an allocation will capture all E&P of the CFCs up to the full amount of the debt obligation, and is based on a tax attribute (E&P) that is central to the operation of section 956 and is unlikely to produce counterintuitive results. Ultimately, a version of this is reflected in our draft rule.

Once we concluded that allocating debt among CFCs on the basis of positive E&P balances is appropriate, the next question was whether to draw a distinction between E&P that is reflected in a PTI balance, and E&P that has not previously been taxed. This question presents the most challenging policy issues. In this regard, there are three potential regimes that come to mind: (i) a regime where allocations are made first on the basis of PTI balances, and then on the basis of remaining positive E&P, so that a United States shareholder is allowed the full use of all of its PTI balances with respect to all CFC guarantors before any taxable inclusion under section 956 is recognized; (ii) a regime where allocations are made on the basis of positive E&P balances without regard to whether E&P has been included in PTI or not; and (iii) a regime (mentioned in the Preamble as a topic for comments) where allocations are made by reference only to untaxed positive E&P balances (so-called "section 959(c)(3) amounts"), so that PTI cannot be used until all untaxed E&P first has been taken into income.²⁴

Before considering the three alternatives in more detail, it is worth identifying certain aspects of the use and generation of PTI balances that we believe are mandated under subpart F regardless of how a debt obligation ultimately is allocated debt among CFC guarantors. First, we believe that any section 956 inclusions that result from such allocations of a debt obligation must give rise to section 956 PTI, which, unless treated as distributed, will in turn be available to offset further inclusions from that same debt obligation in later years. For example, if a United States shareholder owns two CFCs, each of which has $1000 of positive untaxed E&P, and each of which guarantees a $100 debt obligation of the shareholder, we have assumed that the United States shareholder should not recognize a $100 inclusion year after year until the untaxed E&P of the CFCs is entirely exhausted (in 20 years, under these facts, assuming no other transactions).²⁵ Our conclusion is supported, if not compelled, by the statutory language of section 956, which defines a section 956(a) amount as the excess of investments in United States property over section 956 PTI arising from prior section 956 inclusions.

Second, once debt is allocated to a CFC guarantor, we believe that any PTI balances in respect of that CFC guarantor should be available to offset any section 956 inclusion arising from the allocated debt obligation to the same extent that they would be available under section 959 to offset any other distributions or deemed distributions from that CFC (i.e., inclusions under the allocation scheme should not somehow be able to bypass the basic ordering rules for PTI). This result is supported, if not compelled, by the statutory scheme of section 959. A corollary to that rule is that any section 956 PTI balances created through an allocation of debt to a CFC generally should be available to offset other distributions or deemed distributions from that CFC, as other section 956 PTI balances would be.

Third, we think that debt obligations supported by multiple CFC guarantors should be allocated among the guarantors yearly. A rule that allocated debt obligations among CFCs only once would raise numerous issues. For example, imagine two CFCs, each with positive untaxed E&P of $100 and no PTI provide guarantees of a $1000 debt obligation of a U.S. person. The $1000 debt obligation might be allocated evenly between the two CFC, with $500 to each, but since each CFC has only $100 of PTI, there can only be a total of $200 of section 956 inclusions. If in a subsequent year, a third CFC with $800 of untaxed E&P and no PTI provides a guarantee of the debt obligation, clearly that CFC's untaxed E&P should be available to support an addition section 956 inclusion with respect to the debt obligation, notwithstanding the fact that the debt was entirely allocated among the original two CFCs in a prior year. Alternatively, even if no additional CFC provided a guarantee, imagine that the first CFC earned an additional $1000 of untaxed E&P while the second CFC earned nothing. The original allocation of $500 of the debt obligation to the first CFC should not prevent its entire amount of untaxed E&P from being available to support a section 956 inclusion. We think that, rather than providing complicated rules describing when debt obligations should and should not be reallocated among CFC guarantors, the best approach is to reallocate yearly.

With those basic points in mind, we can now consider the specific differences among the allocation methodologies discussed above. As an initial matter, we note that, of course, an allocation of debt among CFC guarantors proportionately to all positive E&P balances or positive untaxed E&P balances (the second and third alternatives described above) is most likely to result in larger up-front taxable inclusions than the first alternative, although presumably with the consequence of creating PTI that can be used to offset income at a later date. We believe, however, the first (allocating debt obligations first proportionately to PTI balances, in accordance with the waterfall in section 959) is not only the most consistent with the general framework of subpart F, but also the most workable solution as a technical matter. Any rule that allocates debt obligations in a manner that does not first look to PTI will create complexity, because such an allocation is contrary to the existing statutory framework under sections 956 and 959. Such a rule either will create distortions in the operation of subpart F and/or require very complex mechanics to categorize and trace PTI under rules that would both override the current regime under section 959 and need to be integrated with it. In particular, we think that a rule allocating debt obligations among CFCs in accordance with their untaxed E&P seems illogical in the context of the current rules under sections 956 and 959, which treat distributions as coming first from PTI. The only way that we can see to make such a rule function properly under current law would be to develop what effectively amounts to a separate regulatory regime for deemed distributions arising from multiple-guarantor situations, and then to develop rules coordinating that regime with the current rules under sections 956 and 959 that apply to all other cases.²⁶ Such an undertaking would involve considerable complexity, and the discussion below describes in more detail some of the specific technical issues that we believe would need to be addressed if Treasury and the IRS were to opt for either the second or third allocation regime.

First, any allocation scheme presumably would need to operate over the course of several years because, as discussed above and like the current rules under section 956, it would need to take account of changes in the composition of the CFC guarantor group over time, other investments made by the CFC guarantors in United States property over the course of several years, changes in E&P balances, etc. However, in order to avoid situations where debt is reallocated among CFC guarantors from year-to-year to create noneconomic inclusions, any debt obligation allocation rule should -- at a minimum -- take into account section 956 PTI created by a section 956 inclusion that resulted from the allocation of a debt obligation under that rule to avoid the clearly inappropriate result discussed above of including the same amount over and over again.²⁷

For example, imagine two CFCs guaranteeing a debt obligation of $100 where, in Year 1, each has $100 of E&P, no PTI and no investment in United States property other than the guarantee of the debt and no subpart F income. An allocation scheme based on relative E&P balances accordingly allocates the debt between them evenly, so that each is deemed to have made a distribution of $50. Then in Year 2, one CFC ("CFC 1") accumulates $100 of additional E&P, but the other ("CFC 2") just breaks even. CFC 1 thus has a positive E&P balance of $200, and CFC 2 continues to have a balance of $100 in respect of Year 2, and nothing else has changed relative to Year 1.

In considering how to apply an allocation rule in respect of Year 2, the first issue is whether the United States shareholder gets "credit" for the $100 inclusion from Year 1. Without a mechanism to give such a credit, the $100 debt obligation would presumably be allocated $66.66 to CFC 1 and $33.33 to CFC 2, resulting in additional $16.66 inclusion with respect to CFC 1. If in Year 3, CFC 1 had a current E&P deficit of $100 and CFC 2 had current E&P of $100 (resulting in CFC 1 and CFC 2 having total current and accumulated E&P of $100 and $200, respectively), presumably there would be yet another inclusion as the debt was allocated now $33.33 to CFC 1 and $66.66 to CFC 2. Unless the allocation regime is going to require duplicative inclusions in later years, despite the fact that the United States shareholder has already recognized a section 956 inclusion equal to the face amount of the debt, there must be some mechanism in place to mark the fact of the prior year's inclusion and prevent a portion of a debt obligation that already gave rise to a taxable section 956 inclusion with respect to one CFC from being reallocated to another CFC to result in a duplicative inclusion in the absence of a transaction or event such as a sale of the first CFC. The first alternative described above (allocating debt obligations first in accordance with PTI) provides such a mechanism automatically. Under one of the other alternatives, if a mechanism takes into account the prior $100 inclusion in allocating debt obligations, but otherwise ignores PTI balances, then it becomes necessary to create and track a new, separate category of PTI. For purposes of discussion, we refer to such a hypothetical category as "Multiple Guarantee Section 956 PTI." Once one goes down that path of creating Multiple Guarantee Section 956 PTI, it is necessary to consider a host of technical issues, such as:

As discussed in more detail below, we believe that the issues discussed above can be addressed most logically and coherently by adopting the first approach discussed above (allocation to achieve full recovery of PTI before recognition of taxable income), and our illustrative draft regulation provides an example of how that might be achieved. We also note that, as a policy matter, we are not as troubled as the authors of the Preamble and certain representatives of the IRS appear to be²⁹ by the prospect of taxpayers being allowed to access PTI fully before recognizing additional taxable income under section 956. Although we acknowledge that taxpayers could mitigate the effect of section 956(d), for example, by adding "high PTI" CFCs to a guarantor group as a way of taking advantage of the CFCs' PTI balances, we would note that current law already allows taxpayers the ability to repatriate cash from CFCs on a tax-free basis up to the amount of the relevant PTI balances -- and to choose to take distributions from CFCs with PTI before looking to those without PTI. Accordingly, some of our members do not necessarily view the use of PTI to enable a deemed distribution from a CFC as very troubling.

We acknowledge, of course, that is impossible to predict all of the different types of tax planning that might develop from the adoption of a rule. For example, our proposed rule would allow a taxpayer to benefit from the credit support of low-PTI CFCs if high-PTI CFCs also provided guarantees. We think it unlikely, however, that taxpayers generally would seek to create high-PTI CFC's for such purposes, since that strategy would entail the recognition of income first in order to prevent the recognition of a similar amount of income at a later point. This point is reinforced by the observation, again, that taxpayers who find themselves unwillingly facing multiple section 956 inclusions in respect of the same debt obligation largely could have avoided section 956 inclusions altogether through a pledge of less than 66 2/3 percent of the CFC stock. We are not sure if there is any use of PTI that could arise under our draft proposed rule that could not be achieved easily under current law with some minimal level of tax planning. In any event, the use of all available PTI to shelter section 956 income in the case of multiple CFC guarantors accordingly does not seem to present a situation where the opportunities for abuse are great. We therefore would be reluctant to recommend that Treasury and the IRS undertake the considerable task of developing fully the concept of Multiple Guarantee Section 956 PTI in regulations to address potential abuses arising from the privileged status of PTI (which, of course, is generally mandated by the ordering rule of section 959).

For the reasons discussed above, we recommend a rule under which debt obligations are allocated among CFC guarantors first in proportion to their relative PTI balances, and then in proportion to untaxed positive E&P balances. Our draft proposed regulatory language has such an allocation scheme at its core. Under such a scheme, there would be no additional section 956 inclusion in Year 2 in the example above, because the debt would be allocated between CFC 1 and CFC 2 equally in Year 2, in proportion to their section 956 PTI balances created in Year 1. Under such a rule, the only way in which there could be additional section 956 inclusions after Year 1 would be if the PTI balances of the guarantor group were reduced (e.g., by a cash distribution allocable against PTI, or if one the CFCs were to be sold).

Our proposed rule achieves the first two objectives by positing a hypothetical entity with a PTI balance and positive E&P balance equal to the sum of all positive E&P balances in respect of the individual CFC guarantors that have positive E&P balances. That single hypothetical entity is then treated as the sole guarantor of all of the obligations of U.S. persons that are guaranteed by more than one CFC of the same United States shareholder. In addition to achieving the objectives of aggregating PTI and E&P, the hypothetical entity then allows for clear rules as to how those attributes are to be used: i.e., they lead to the creation of a section 956(a) amount and a section 956 inclusion in the same manner as if a CFC with those aggregate attributes had guaranteed the relevant debt obligations. In this way, the hypothetical entity mimics the standard application of section 956 to the greatest extent possible and thus simplifies the application of what might otherwise be very complicated rules.

In considering how best to calculate the amount described in section 956(a)(1)(A) for the hypothetical combined entity, we first considered using some sort of weighted average of values of the investment in United States property resulting from the guaranty provided by each individual CFC guarantor to determine the United States shareholder's "pro rata" share of the entity's deemed investment in United States property, as well as the entity's quarterly positions that need themselves to be averaged. Although such an approach is more nuanced (and arguably more economically accurate) than the approach we adopted for our draft language, the more nuanced approach would introduce considerable complexity. We do not believe that such complexity is justified. In this regard, we note that the issue relating to section 956(a)(1)(A) is unlikely to be relevant except on rare occasions, since the vast majority of cases implementing our rule are going to involve CFC guarantors that are all wholly owned within a large corporate group and that all guarantee the same debt obligation simultaneously.

Instead of a weighted average approach, therefore, we recommend using the largest value calculated for the relevant United States shareholder under section 956(a)(1)(A) in respect of all of the CFC guarantors. For example, consider the case where a United States shareholder owns 70% of CFC1, which guarantees an obligation for the entire year, and 100% of CFC 2, which guarantees the obligation only during three calendar quarters. The largest value calculated under section 956(a)(1)(A) would be 75% of the obligation (since 100% of the obligation for three of four quarters is more than 70% of the obligation for four of four quarters).

The use of the largest has the virtue of simplicity. However, by looking simply to the largest value, our rule could generate a section 956 inclusion that is greater than what might arise under current law. For example, if CFC2 has an E&P deficit and therefore cannot pay a deemed dividend under section 956, then the 75% value determined by reference to CFC2 would be used to determine a section 956 inclusion that draws on E&P residing in CFC1. The inclusion under this rule would be greater than the inclusion under current law, where the lower 70% percentage would have been used to calculate amounts deemed paid by CFC1. Because we do not intend for our proposed rule to generate larger section 956 inclusions than those that would be generated under current law, and would not consider such a result to be appropriate, under our suggested approach the section 956 amount generated in respect of the hypothetical combined entity would by its terms function merely as a cap on the United States shareholder's inclusion under section 956 with respect to a CFC, and accordingly will not produce greater inclusions than would current law.

Once the overall section 956 amount is determined by reference to the hypothetical combined entity (so that the use of PTI balances and positive E&P balances is maximized), our rule then allocates that overall section 956 amount among the individual CFC guarantors through a waterfall. First, the section 956(a) amount is allocated among the CFC guarantors on a pro rata basis in accordance with their positive subpart F income PTI balances³⁰, to the extent thereof, and second, any remaining amount of the inclusion is allocated among the CFC guarantors pro rata in accordance with their remaining positive E&P balances. Each CFC guarantor is then treated for all purposes (e.g., utilization of foreign tax credits, calculation of new PTI balances and other relevant tax attributes) as if it had a section 956(a) amount equal to its allocation under the waterfall. Because of this treatment, the tax attributes resulting from the hypothetical combined entity's section 956(a) amount are localized within specific CFCs, where they can be tracked and taken account of in subsequent years along with the CFC's other attributes. If one CFC guarantor ceases to guarantee the relevant debt, it nonetheless will retain the tax attributes (PTI balances etc.) reflecting its share of the section 956 inclusion from the taxable years in which it was a guarantor. Because those attributes are now specific to that CFC, however, they would not be taken into account in subsequent years, for example, in the new calculation of section 956 inclusions for the remaining CFC guarantors by reference to a new hypothetical combined entity that no longer includes that CFC.

At the beginning of Year 2, CFC 3 is sold and ceases to be a CFC with respect to the U.S. shareholder. The Year 1 section 956(a) amount allocated to CFC3 is reflected in the U.S. shareholder's basis in CFC3, so that on disposition the taxpayer does not suffer an additional tax on the $30 already included in income. CFC 1 and CFC 2 continue to guarantee the $90 borrowing for all of Year 2, and at the end of Year 2 have section 956 PTI balances of $30 each and total positive E&P of $100 each. Under our proposed rule, the hypothetical combined entity's section 956(a) amount for Year 2 is $30, calculated as the lesser of (1) the excess of the $90 borrowing over the aggregate $60 of section 959(c)(1)(A) amounts, and (2) the aggregate $140 of applicable earnings of the two CFCs. This $30 hypothetical combined entity section 956(a) amount is allocated among the CFCs first in accordance with subpart F income PTI, and then according to other positive E&P, resulting in $15 being allocated to each CFC. The $15 allocated to each CFC is a section 956(a) amount for all purposes of the Code, resulting in an inclusion under section 951(a)(1)(B) and the creation of an additional $15 of section 959(c)(1)(A) PTI with respect to each CFC.

Finally, at the beginning of Year 3, the U.S. shareholder acquires an additional CFC, CFC4, which guarantees the $90 borrowing, and CFC1, CFC2 and CFC4 guarantee the borrowing for all of Year 3. During Year 3, CFC2 distributes $20, which is treated under section 959 as a distribution out of CFC2's $45 balance of section 959(c)(1)(A) PTI, and CFC4 earns $8 of subpart F income. At the end of Year 3, CFC 1 has $45 of section 956 PTI and $100 of total positive E&P, CFC2 has $25 of section 956 PTI and $80 of total positive E&P, and CFC4 has $8 of subpart F income PTI and $118 of total positive E&P. Under our proposed rule, the hypothetical combined entity's section 956(a) amount for Year 3 is $20, calculated as the lesser of (1) the excess of the $90 borrowing over the aggregate $70 of section 959(c)(1)(A) amounts, and (2) the aggregate $228 of applicable earnings of the three CFCs. This $20 hypothetical combined entity section 956(a) amount is allocated among the CFCs first in accordance with subpart F income PTI, with $8 being allocated to CFC4, and then according to other positive E&P, resulting in $3 being allocated to each of CFC1 and CFC2 and an additional $6 being allocated to CFC4 (CFC1 and CFC2 each has $55 of untaxed E&P, equal to total positive E&P of $100 and $80, respectively, in excess of section 956 PTI balances of $45 and $25, respectively and CFC4 has $110 of untaxed E&P, equal to total positive E&P of $118 in excess of section 956 PTI balance of $8). The amount allocated to each CFC is a section 956(a) amount for all purposes of the Code, resulting in an inclusion under section 951(a)(1)(B) of $3 each by CFC1 and CFC2 and $6 by CFC4 (and reclassification of $8 of subpart F income PTI, in the case of CFC4) and the creation of additional section 959(c)(1)(A) PTI with respect to each CFC.

V. MECHANICAL OPERATION OF THE PROPOSED RULE.

This Part V provides a more technical discussion of the proposed rule contained in the Appendix.

In order to determine the section 956(a) amount for a given United States shareholder in respect of a section 956(d) group obligation, our illustrative draft Treasury regulation section 1.956-1(h) looks to each individual CFC member of the section 956(d) group, and limits each such CFC's section 956(a) amount in respect of the United States shareholder to the sum of:

The "hypothetical combined entity amount," in turn, is determined in a calculation analogous to the calculation in section 956(a), but based on the combined attributes of all of the CFCs in the section 956(d) group remaining after excluding attributes utilized in determining the standalone entity section 956 amount of each CFC (i.e. attributes used in determining the section 956(a) amount as if there was no section 956(d) group obligation). Specifically, the hypothetical combined entity amount is determined by reference to the excess of a United States shareholder's share of section 956(d) group obligations over the section 956(d) group's combined positive section 956 PTI balance. This combined section 956 PTI balance is determined by adding the section 956 PTI balances of the individual CFCs, but only after taking into account at the level of each individual CFC the application of section 956 PTI balances to any section 956(a) amounts not attributable to the section 956(d) group obligation. Assuming that the United States shareholder's share of the section 956(d) group obligation in fact is greater than the combined available section 956 PTI balance, the proposed regulatory definition caps the hypothetical combined entity amount at the level of the group's applicable earnings (determined as the sum of the positive amount of applicable earnings, as determined in accordance with section 956(b)(1) of all members of the section 956(d) group). The rules contain a provision requiring that the applicable earnings of a CFC be reduced by amounts used to calculate a section 956(a) amount with respect to a standalone investment in United States property, so that the same E&P amount cannot support a section 956(a) amount with respect to both a section 956(d) group obligation and another standalone investment by the CFC in United States property.

In drafting the definition of "hypothetical combined entity amount," we took into account of the fact that, by virtue of the operation of section 956(a), a United States shareholder's share of a section 956(d) group obligation, at least in the case of a single CFC guarantor, is a function of: (i) the United States shareholder's equity stake in the relevant CFC (since section 956 amounts are determined by reference to a United States shareholder's pro rata share of an investment in United States property based on equity ownership), and (ii) the times when the CFC became, and ceased to be, a guarantor supporting the obligation (since the section 956 amount is the result of average quarter-end balances, and the guarantee might not be in place at the end of each quarter in a given taxable year). Rather than providing a complicated averaging or weighting approach in order to deal with this case or multiple CFC guarantors, our recommended Treasury regulation section 1.956-1(h)(3)(B) merely looks to the CFC for each section 956(d) group obligation that produces the largest share and attributes that share to the United States shareholder as an investment supported by the entire section 956(d) group.

Once the hypothetical combined entity amount is determined, each CFC's share of that amount is determined under a "waterfall" in the following manner: First the hypothetical combined entity amount is allocated among the members of the section 956(d) group in an amount up to their combined subpart F income PTI balances, on a pro rata basis by reference to such balances. This allocation has the effect of maximizing the subpart F income PTI balance for the group that are available to offset inclusions. Then remaining amounts are allocated among the members in proportion to their positive E&P amounts in excess of their subpart F income PTI balances. By allocating the combined entity amount first by reference to subpart F income PTI balances, the rule seeks to maximize the use of such balances to offset the section 956(a) amount in calculating the United States shareholder's taxable income. In each case, the subpart F income PTI balances and the positive E&P amounts are those amounts after any reductions necessary to take account of a CFC's standalone section 956(a) amounts (i.e., section 956(a) amounts attributable to investments in United States property other than those treated as held by such CFC by way of section 956(d)).

As an initial matter, and as discussed above, the proposed rule uses the term "section 956(d) group" to refer to all of the CFCs of the United States shareholders that each guarantee a particular section 956(d) group obligations (or multiple section 956(d) group obligations, if all such section 956(d) group obligations have the same CFCs as guarantors). Situations where a United States shareholder has one section 956(d) group or multiple section 956(d) groups that do not overlap are straightforward -- there is only one section 956(d) group to which a CFCs positive E&P can be allocated. The more complex situation arises when a United States shareholder has more than one section 956(d) group that partially overlap, so the positive E&P of one or more CFCs needs to be divided between two section 956(d) groups (i.e., between two sets of one or more debt obligations) to determine the United States shareholder's total section 956 inclusion.

Although we initially set out to draft a precise mechanism for reflecting the principles outlined above, the exercise proved to create a set of proposed rules that were overly complex. Given that the situation these rules are attempting to address is, we think, less commonplace, the proposed rule instead provides that an allocation of earnings and profits and PTI among section 956(d) groups shall be respected so long as it is reasonable. The proposed rule then provides a series of principles that establish the boundaries of what is reasonable: earnings and profits and PTI should not be double counted (i.e., should not be used by more than one section 956(d) group); earnings and profits and PTI of a CFC should not be used by a section 956(d) group of which that CFC is not a member; the amount of earnings and profits utilized by an apportionment to a section 956(d) group should not be in excess of the amount of obligations treated as held by that group; earnings and profits and PTI should be apportioned to section 956(d) groups in a manner that minimizes the amount of "leftover" earnings and profits and PTI (i.e., taxpayers get the full benefit of available PTI to offset income inclusions, but non-PTI earnings and profits should be utilized to the extent possible).

² Unless indicated otherwise, all section references in this report are either to the Code or to Treasury regulations promulgated thereunder.

³See Notice of Proposed Rulemaking, Fed. Reg. Vol. 80, No. 170, p. 53058 (September 2, 2015).

As a general matter, section 957 defines a "controlled foreign corporation" as a foreign corporation, more than 50 percent of whose equity (measured by vote or value) is owned by "United States shareholders". "United States shareholder" in turn is defined in section 951(b) generally as a U.S. person that owns directly, indirectly or constructively 10 percent or more of the total combined voting power of the corporation's stock.

⁴ Subpart F refers to subpart F of Part III of subchapter N of the Code.

⁵ A United States shareholder is defined in section 951(b) as a U.S. person who owns or is considered as owning 10 percent or more of the total combined voting power of all classes of stock entitled to vote. For this purpose, U.S. person generally includes persons defined as such by section 7701(a)(30) (i.e., U.S. individual citizen or resident alien, domestic corporation, domestic partnership). Ownership means direct and indirect ownership (as defined by section 958(a)(1)-(2)) as well as constructive ownership, as determined under section 318(a) (provided that for this purpose, (i) stock owned by a nonresident alien individual is not considered as constructively owned by a resident alien or individual; (ii) if a partnership, estate, trust or corporation owns directly or indirectly more than 50 percent of voting power of a corporation it shall be considered as owning 100 percent of the voting power; and (iii) if 10 percent or more of a corporation is owned, shareholders are treated as owning their proportionate share of the corporation's assets).

⁶ Subpart F income is defined in section 952. It includes, for any CFC, the sum of such CFC's insurance income (as defined in section 953), foreign base company income (as defined in section 954) and income attributable to certain jurisdictions with which the U.S. has designated as having poor trade relations, pays illegal bribes or kickbacks or participates in any international boycott. Section 952(a). Subpart F income of any CFC for a taxable year is limited to such CFC's earnings and profits for that taxable year and does not include any U.S.-source income that is effectively connected with the conduct of a trade or business in the U.S. Section 952(b)-(c). Foreign base company income, defined in section 954, includes foreign personal holding company income, foreign base company sales income, foreign base company services income and foreign base company oil related income. Section 954(a). In order for Subpart F income of a CFC to give rise to an income inclusion for a United States shareholder thereof, the CFC must be a CFC for an uninterrupted period of at least 30 days during the taxable year. Section 951(a)(1).

⁷ Section 956(c)(1).

⁸ The legislative intent for the enactment of section 956 was to ensure that U.S. shareholders of certain foreign corporations did not enjoy tax deferral on foreign earnings invested in United States property to the extent those earnings were not otherwise taken into account on a current basis under subpart F. See S. Rep. No. 1881, 87th Cong., 2d Sess. 80, 87-88 (1962), 1962-3 C.B. 707 at 794 ("Generally, earnings brought back to the United States are taxed to the shareholders on the grounds that this is substantially the equivalent of a dividend being paid to them.").

⁹ Section 956(a)(1)-(2) (calculation of the section 956 amount) and section 956(b)(1) (definition of applicable earnings).

¹⁰ Section 951(a)(1)(A) (calculation of subpart F income inclusions) and section 959(a)(1).

¹¹See Section 956(a)(1)(B) and section 959(f)(1).

¹² Section 959(c)(1), 959(f)(1).

¹³See Sections 951(a)(1)(B), 959(a)(2), (f).

¹⁴ Section 959(c)(1)(A).

¹⁵ Section 959(c)(2).

¹⁶See Section 959(c).

¹⁷ It is worth noting, however, that the construct under section 956(d) may create anomalous results where the United States shareholder/borrower owns less than 100 percent of the CFC guarantor's stock. As discussed below, section 956(d) treats the CFC guarantor as holding the guaranteed debt obligation, and thus as having made an investment in United States property. However, since the United States shareholder is required to recognize only the pro rata portion of the investment based on its ownership of the CFC's stock, it will not recognize the full amount of the borrowing as a section 956 inclusion. Perhaps even more anomalous, some portion of the section 956 amount might instead be recognized by another unrelated United States shareholder in the CFC that neither borrowed nor benefitted from the borrowing.

¹⁸ Treasury regulations section 1.956-1(e)(2) ("For purposes of this section the amount taken into account with respect to any pledge or guarantee described in paragraph (c)(1) of § 1.956-2 shall be the unpaid principal amount on the applicable determination date of the obligation with respect to which the controlled foreign corporation is a pledgor or guarantor.")

¹⁹ 1976-1 C.B. 204.

²⁰Ludwig v. Commissioner, 68 T.C. 979 (1977).

²¹See Id.

²² It should, however, be possible for section 956 amounts over time to aggregate to more than the amount of the guaranteed debt in cases where an inclusion gives rise to a PTI balance that is later reduced under section 959 (e.g., through a repatriation of cash). That outcome is no different than what would occur in the case of only a single CFC guarantor.

²³ Generally guarantors would have "joint and several liability" in respect of their obligations under the guarantee, so that in the event the borrower defaulted, the lenders could collect the full amount of the borrowing from any one, several or all of the guarantors. In the event one guarantor satisfies the entire liability to the lender, it may or may not have a right of contribution against the other guarantors in the credit support group so that each would ultimately bear an economic burden in respect of the guarantee. If each guarantor has only "several liability", then it would be liable for only that portion of the obligation guaranteed by it.

²⁴ Of course, there are multiple other regimes that one could devise by coming up with hybrids of these regimes, but we believe that the three regimes described above are sufficient to inform the policy discussion.

²⁵ It would not be appropriate to replace the existing problems arising from multiple CFCs providing guarantees of a single debt obligation with a Nietzschean eternal recurrence of taxable income.

²⁶ In this regard, certain of our members expressed a concern that the language in the Preamble could suggest that the IRS and Treasury are considering a rule diverging even further from the basic operation of subpart F by requiring a deemed repatriation of untaxed E&P from a CFC before any deemed repatriation of PTI from that CFC. In other words, once debt is allocated to a CFC, such a rule effectively would ignore the ordering rule contained in section 959 and create taxable income to the extent of a CFC's untaxed E&P, regardless of what PTI balances might exist. All of our members agree that such a rule would be inappropriate for a number of reasons. First, such a rule would create as much, or more, complexity as a rule that merely allocates debt in accordance with untaxed E&P. We do not think it makes sense to create what amounts to an alternative subpart F regime simply to address what appears to be an oversight in section 956(d). Second, and more important, however, section 959 creates a clear ordering rule. We do not believe it is appropriate to require taxpayers to treat multiple-guarantor dividends under section 956(d) as somehow exempt from that clear (and clearly intended) statutory rule.

²⁷See text accompanying note 25.

²⁸ Although, as discussed above, we believe that all types of PTI created under subpart F must generally be made available in accordance with the terms of section 956 and 959, it is possible to imagine a number of rules designed to match certain types of PTI with certain types of income consistently with that principle, and the questions in the text above are some that are created by such an undertaking.

²⁹ Andrew Velarde, Treasury Open to Goodwill Exception When Abuse is Unlikely, Tax Notes, November 5, 2015 (quoting Douglas Poms, Treasury department international tax counsel, "We want a fair rule so that there is one inclusion of the unpaid principal amount of the obligation, but it is an effective inclusion, not one that you could easily manipulate to use all of a particular CFC's previously taxed income upfront.").

³⁰ The draft regulatory language refers only to subpart F income PTI balance at this stage because section 956 PTI has already been taken account of in the calculation of the section 956 amount.

(1) In general. This paragraph applies only in the case of a United States shareholder of more than one controlled foreign corporation, each of which is a member of a section 956(d) group (as defined below) with respect to such United States shareholder. The provisions of this paragraph shall be applied in a consistent manner for the relevant taxable period of determination, taking into account the principles of section 1.956-1(d) and the purpose of this paragraph.

(3) Hypothetical combined entity amount.

(4) Controlled foreign corporation's share of the hypothetical combined entity amount. For any taxable year, the share of the hypothetical combined entity amount for each controlled foreign corporation that is a member of the section 956(d) group to which the hypothetical combined entity amount relates equals the sum of --

(5) Exclusion of earnings and profits taken into account with respect to the standalone entity section 956 amount of a controlled foreign corporation. For any taxable year, any earnings and profits of a controlled foreign corporation with respect to a United States shareholder described in subparagraphs (A), (B) or (C) of this subparagraph (h)(5) shall be excluded for purposes of determining a hypothetical combined entity amount under subparagraph (h)(3)(A), determining a controlled foreign corporation's share of the hypothetical combined entity amount under subparagraph (h)(4), or allocating earnings and profits among section 956(d) groups under subparagraph (h)(7)(B).

Example (1). A is a United States shareholder and direct owner of 100% of the only class of stock of FC1 and 80% of the only class of stock of FC2, each of which is a controlled foreign corporation for the entire period involved. A borrows $100,000 from a bank at the beginning of Year 1, and FC2 provides a guarantee of the loan. During the third quarter of Year 1, A borrows an additional $50,000 from the bank, also guaranteed by FC2, and in connection with the additional borrowing FC1 guarantees the entire $150,000 of outstanding borrowings. Neither FC1 nor FC2 holds any other United States property, or makes any distribution during Year 1.

FC1 and FC2 constitute a "section 956(d) group" within the meaning of this paragraph, because each is treated as holding the same obligations of a U.S. person under section 956(d) during Year 1, and each of the $100,000 borrowing and the $50,000 borrowing is treated as a "section 956(d) group obligation." For purposes of determining A's share of the section 956(d) group obligations under subparagraph (h)(3)(B), A's pro rata share of each obligation (calculated in accordance with the principles of section 956(a)(1)(A)) is determined separately with respect to each of FC1 and FC2. A's largest pro rata share of each obligation is identified and A's share of section 956(d) group obligations equals the sum of A's largest pro rata share in respect of each obligation.

FC1 was treated under section 956(d) as holding both the $100,000 and $50,000 obligation from the third quarter of Year 1, so its average amount of each obligation held at the close of each quarter of Year 1 is $50,000 and $25,000, respectively. A holds 100% of the sole class of stock of FC1, so its pro rata share of FC1's average quarterly balances is 100% of such balances, or $50,000 and $25,000, respectively.

FC2 was treated under section 956(d) as holding the $100,000 obligation from the first quarter of Year 1 and the $50,000 obligation from the third quarter of Year 1, so the average amount of each obligation held by FC2 at the close of each quarter of Year 1 is $100,000 and $25,000, respectively. A holds 80% of the sole class of stock of FC2, so its pro rata share of FC2's average quarterly balances is $80,000 and $20,000, respectively.

A's largest pro rata share of the $100,000 obligation is $80,000 (its pro rata share with respect to FC2, which is greater than its pro rata share with respect to FC1). A's largest pro rata share of the $50,000 obligation is $25,000 (its pro rata share with respect to FC1, since $25,000 is greater than $20,000). As a result, A's share of the section 956(d) group obligations for Year 1 is $105,000.

Because FC1 and FC2 did not hold any United States property in Year 1 other than the section 956(d) group obligations, none of their earnings and profits are excluded from the calculation of the "hypothetical combined entity amount" of the section 956(d) group.

Further assume that the section 959(c) amount of FC1 and FC2 with respect to A, after taking into account any section 951(a)(1)(A) amounts for Year 1, are as follows:

                                                        FC1      FC2

 (i)     A's share of the section 956(d)

                                                        FC1      FC2

Example (2). B is a United States shareholder and direct owner of 100% of the only class of stock of each of FC3 and FC4, each of which is a controlled foreign corporation for the entire period involved. B borrows $100,000 from a bank at the beginning of Year 1, and FC3 and FC4 provide guarantees of the loan. B borrows an additional $50,000 from the bank during the third quarter of Year 1, but this additional borrowing is guaranteed only by FC3. Neither FC3 nor FC4 holds any other United States property nor makes any distributions during Year 1.

FC3 and FC4 constitute a "section 956(d) group" within the meaning of this paragraph, because each is treated as holding the same section 956(d) obligation during Year 1. The $100,000 borrowing is treated as a "section 956(d) group obligation" because it is treated as held by more than one member of the section 956(d) group during Year 1. The $50,000 borrowing is not treated as a section 956(d) group obligation, because it is treated as owned only by FC3.

For purposes of determining B's share of the section 956(d) group obligation under subparagraph (h)(3)(B), B's pro rata share of the section 956(d) obligation (calculated in accordance with the principles of section 956(a)(1)(A)) is determined separately with respect to each of FC3 and FC4. B's largest pro rata share of the obligation is identified and B's share of the section 956(d) group obligations equals the sum of B's largest pro rata share in respect of the obligation. Because both FC3 and FC4 held the section 956(d) group obligation for all of Year 1 and B owns 100% of the single class of stock of each of FC3 and FC4, B's share of the section 956(d) group obligations equals the face amount of such obligation, or $100,000. Further assume that the amounts of earnings and profits of FC3 and FC4 with respect to B, after taking into account any section 951(a)(1)(A) amounts for Year 1, are as follows:

                                                        FC3      FC4

                                                        FC3      FC4

 (i)     B's share of the section 956(d) group

                                                        FC3      FC4

(7) Determination in the case of section 956(d) group obligations not treated as held by all members of the section 956(d) group.

Example. C is a United States shareholder and direct owner of 100% of the only class of stock of each of FC5, FC6, and FC7, each of which is a controlled foreign corporation for the entire period involved. C borrows $100,000 from bank X at the beginning of Year 1, and FC5 and FC6 provide guarantees of the loan. At the same time, C borrows $50,000 from bank Y, and FC5, FC6 and FC7 provide guarantees of the loan. None FC5, FC6 or FC7 hold any other United States property, or makes any distribution during Year 1. The borrowings and their guarantors are summarized as follows:

 $100,000 borrowing                                FC5 and FC6

For purposes of determining C's share of the section 956(d) group obligations under subparagraph (h)(3)(B), C's pro rata share of each obligation (calculated in accordance with the principles of section 956(a)(1)(A)) is determined separately with respect to each of FC5, FC6 and FC7.

FC5 and FC6 were treated under section 956(d) as holding the $100,000 during all of Year 1, so both FC5 and FC6's average amount of the obligation held at the close of each quarter of Year 1 is $100,000. C holds 100% of the sole class of stock of each of FC5 and FC6, so its pro rata share of FC5 and FC6's average quarterly balances is 100% of such balances, or $100,000.

FC5, FC6 and FC7 were treated under section 956(d) as holding the $50,000 during all of Year 1, so each of FC5, FC6 and FC7's average amount of the obligation held at the close of each quarter of Year 1 is $50,000. C holds 100% of the sole class of stock of each of FC5, FC6 and FC7, so its pro rata share of FC5, FC6 and FC7's average quarterly balances is 100% of such balances, or $50,000.

As a result, C's share of the section 956(d) group obligations is the sum of (1) C's pro rata share of the $100,000 obligation with respect to either FC5 or FC6 (whichever is larger), or $100,000 in this case, and (2) C's pro rata share of the $50,000 obligation with respect to either FC5, FC6 or FC7 (whichever is largest), or $50,000 in this case. C's share of the section 956(d) group obligations for Year 1 is therefore $150,000.

Further assume that the amounts of earnings and profits of FC5, FC6 and FC7 with respect to C after taking into account any section 951(a)(1)(A) amounts for Year 1, are as follows:

                                               FC5      FC6      FC7

Under subparagraph (h)(7)(B)(ii), the earnings and profits of a controlled foreign corporation may not be allocated to a section 956(d) group of which such controlled foreign corporation is not a member. As a result, no earnings and profits of FC7 are allocated to the FC5, FC6 group with respect to the $100,000 obligation. As a result, at most $90,000 of earnings and profits (the total combined earnings and profits of FC5 and FC6) can be allocated to the FC5, FC6 group with respect to the $100,000 obligation. Thus the section 956(d) group obligation of such group will exceed the amount of earnings and profits allocated to such group, within the meaning of subparagraph (7)(B)(iii). No such excess exists in respect of the group with respect to the $50,000 obligation, because all $170,000 of FC7's earnings and profits may be allocated to such group. In this case, the "aggregate excess with respect to all section 956(d) groups," within the meaning of subparagraph (7)(B)(iii), is $10,000, or the excess with respect to the FC5, FC6 group. This aggregate excess of $10,000, caused by allocating only $90,000 of earnings and profits to the FC6, FC7 group, is permissible under subparagraph (7)(B)(iii), because there is no alternative allocation that creates a smaller aggregate excess. Alternative allocations that create a larger aggregate excess (e.g., allocating only $80,000 of earnings and profits to the FC5, FC6group and allocating $180,000 of earnings and profits to the FC5, FC6, FC7 group, resulting in a $20,000 aggregate excess) are impermissible under subparagraph (7)(B)(iii).

The amounts of earnings and profits of FC5, FC6 and FC7 with respect to C are as follows:

                                               FC5      FC6      FC7

 With respect to the group holding the $100,000 obligation

                                               FC5      FC6      FC7

                                               FC5      FC6      FC7