Tax Court Approves Tax-Free Payment of Outbound Guarantee Fees
by Michael J. Miller
Published: August 13, 2010
Source: BNA Tax Management - International Journal
Originally published in: BNA Tax Management—International Journal August 13, 2010 Tax Court Approves Tax-Free Payment of Outbound Guarantee Fees By: Michael J. Miller and Libin Zhang Recently, in Container Corporation v. Comr., the Tax Court held that guarantee payments are sourced by reference to the “location” of the guarantor and, therefore, are foreign-source income when paid by a U.S. subsidiary to a foreign affiliate. Thus, the Tax Court apparently approved a technique for using guarantees to strip income from the United States to foreign affiliates on a tax-free basis. However, as discussed below, the ability of foreign-based multinationals to take advantage of this decision is doubtful. FACTS The key facts in the case are straightforward. Container Corporation, the taxpayer, is a successor in interest to Vitro S.A. (“Vitro”), a Mexican bottle-making company with a U.S. subsidiary called Vitro International Corporation (“International”). In 1991, International issued senior notes to a consortium of lenders, and used the proceeds to refinance various debts incurred in acquiring various American competitors. Vitro guaranteed the notes in order to provide credit support and to make the debt more marketable. Vitro charged a guarantee fee of 1.5% of the outstanding principal balance of the notes per year. International paid the fees to Vitro, in amounts ranging from $2.3 million in 1992 to $1.9 million in 1993 to $2.5 million in 1994. International did not withhold any U.S. tax on the fee payments. The IRS asserted that International should have withheld 30% of its guarantee fee payments. Section 881(a) of the Code imposes a 30% tax on fixed or determinable annual or periodical income (“FDAP income”) received by a foreign corporation from sources within the United States if that income is not effectively connected with the conduct of a trade or business within the United States. Taxes owed under Search7RH 881(a) generally must be withheld by the payor at the source under Search7RH 1442(a). The definition of FDAP income is sufficiently broad to encompass virtually all kinds of income other than capital gain, so the parties agreed that the 30% withholding tax would apply imposed if the guarantee fees received by Vitro were U.S.-source. As expected, the IRS argued that the guarantee fees paid by International were U.S.-source, while Vitro argued that they were foreign-source. As indicated above and explained further below, the Tax Court sided with the taxpayer. Tax Court Analysis The Tax Court began the discussion with the well-established proposition that, where there is no sourcing rule in the Code for a particular type of income, it is necessary to source by analogy to a type of income for which a sourcing rule is provided. Thus, the Tax Court stated that if guarantee fees were determined to be neither interest nor payments for services, “we would still need to figure out whether they were more like interest or more like payments for services rendered (or, possibly, some other category of FDAP income that has a specific sourcing rule),” citing as precedence Hunt v. Comr., Howkins v. Comr., and Bank of America v. U.S. The Tax Court then considered whether guarantee fees could be characterized as either interest or fees for services, because an affirmative conclusion would conveniently have obviated the need to look beyond the Code to divine an appropriate sourcing rule. Noting that it agreed with the parties, the Tax Court concluded that the fees were not interest because Vitro’s guarantee was not a loan to International. The analysis of whether Vitro’s guarantee fees could be characterized as compensation for services was a bit more labor-intensive. Preliminarily, the Tax Court stated that Vitro was required to maintain records and supply information to the lenders, but noted (with implicit approval) the IRS’s argument that these services were not the “predominant feature” of the guarantee and should therefore be disregarded. More interesting was the taxpayer’s argument that a guarantee is itself a service. Surprisingly, perhaps, the IRS has taken precisely this position on more than one occasion. GCM 38499 (9/19/80), agreeing with a proposed revenue ruling, expressly concluded that “P’s guarantee is the performance of a service for S. . . .” Based on this characterization, the GCM held that the IRS could impute a guarantee fee to P under former Regs. Search7RH1.482-2(b)(1), which applied “[w]here one member of a group of controlled entities performs marketing, managerial, administrative, technical, or other services for the benefit of, or on behalf of another member of the group without charge, or at a charge which is not equal to an arm’s length charge ….” The GCM’s conclusion that guarantees are services was based on an example in the Treasury Regulations that characterized certain activities, including the arranging of financing for a construction project, as collectively constituting services. Also helpful to the taxpayer’s cause was TAM 8508003, which held that guarantee fees are not investment income within the meaning of Search7RH 163(d). When the TAM was issued, Search7RH163(d)(2)(B) defined investment income to include gross income from interest, dividends, rents, and royalties. Based on the conclusion that guarantee fees are not interest, it followed quite easily that such fees did not qualify as investment income. Interestingly, the TAM went further, expressly characterizing guarantee fees as compensation for services: Unlike bona fide interest income, the income here is not a passive return on capital; it is compensation for the act of serving as guarantor of a loan. Compensation for services rendered is not investment income within the meaning of Search7RH163(d)(3)(B). The Tax Court was not particularly impressed by the GCM, and did not even mention the TAM. The opinion does not address the GCM head-on, but seemed to suggest that any weight otherwise accorded to the GCM is offset by Centel Communications v. Comr. The Centel court considered whether three shareholders who initially guaranteed loans to their corporation without compensation recognized compensation income under Search7RH83 when, several years later, they received stock warrants for the guaranteed. As described in the Container Corp. opinion, Centel “held that ‘within the meaning of Search7RH83’ the shareholder had not performed a service.” The Tax Court continued: “[W]ithin the meaning of section 83” is the key. We did characterize the guaranties as “shareholder/investor actions to protect their investment ... [that] as such do not constitute the performance of services.” But we also stressed that our decision turned on a question of fact: whether the shareholders got the warrants in exchange for services rendered as employees or independent contractors. The parties agreed the shareholders weren’t employees, and we found that they were not independent contractors because they were not in the business of guaranteeing loans. We did not hold that providing a guaranty is never a service, and noted that we were analyzing only the language of section 83. [Footnote references omitted.] The Tax Court also seemed to consider it significant that, in a 1995 Field Service Advice, the IRS acknowledged uncertainty regarding the proper treatment of guarantees. An analysis under [section 83] is quite different from an analysis under the sourcing rules, but it nevertheless prompted the Commissioner to rethink his position when the problem came up in the transfer-pricing context again. This time he reasoned that The Centel decision increases the litigating hazards .... However, we do not read this case as contradicting the position of the Service as established in *** G.C.M. 38499. Guarantees do not fit comfortably within normal tax law concepts in a number of areas and, consequently, there are substantial arguments that can be made against any possible analysis of guarantees. *** [1995 WL 1918236 (IRS FSA May 1, 1995).] In light of the Centel decision and the subsequent Field Service Advice, the Tax Court in Container Corp. apparently decided that the whole transfer pricing area was too muddled to provide any precedential or analytical value and that it need not worry very much about the GCM. “All we can conclude from this detour through transfer-pricing law is that it will not help us reach a reasonable conclusion on whether guaranties are services under section 861.” Having decided that all else failed, the Tax Court resorted to consulting the dictionary, and there it found the guidance it sought: So we’ll fall back on the dictionary. The common meaning of “labor or personal services” implies the continuous use of human capital, “as opposed to the salable product of the person’s skill.” Under this definition, we find that Container failed to prove that Corporativo performed sufficient “labor or personal services” to justify the $6 million International paid in guaranty fees over three years. Container presented very little evidence about the specific acts Corporativo performed and how much time it took to perform them. For example, Container’s posttrial brief explains that the Guaranty agreement required Vitro to “take certain actions, confirm certain facts, provide certain information, and create and supply certain documents.” The Guaranty agreement required only minimal accountings and reporting to the note purchasers. In any event, the fees were not tied to the amount of work that Vitro did, but to the amount of the outstanding principal that Vitro was standing behind. This leads us to hold that International did not pay the guaranty fees to Vitro as compensation for services. The value of Vitro’s guaranty stems “from a promise made and not from an intellectual or manual skill applied.” Bank of Am., 47 AFTR 2d at 81-657. [Footnote reference omitted.] Thus, the court concluded that guarantees are not services, and proceeded to embark upon the task of sourcing by analogy. The Tax Court began by examining the analysis of financial payments in Bank of America v. U.S. Bank of America was paid commissions for substituting its own credit for another bank’s credit in acceptance or confirmation of letters of credit, and the Court of Claims held that the commissions were analogous to interest. The Court in Container Corp. distinguished Bank of America and held that the use of Vitro’s credit in a guarantee was different from the use of a bank’s credit in acceptance and confirmation transactions. Bank of America had assumed an unqualified primary obligation to pay the seller, in effect making a short-term loan where the commissions approximated interest. In contrast, the court observed, a guarantor like Vitro did not substitute International’s credit, but merely augmented it. Vitro’s obligations were entirely secondary and did not necessarily increase Vitro’s liabilities or put its assets at risk, as Vitro had legal recourse against International in case the guarantee was triggered. Accordingly, the Tax Court concluded that Vitro’s guarantee fees did not look much like interest. As a fallback, the IRS argued that, if guarantees are unlike loans because the guarantor does not immediately advance funds to the other party, perhaps guarantee fees are “like interest in some broader sense” under Howkins. In Howkins, the source question arose when a U.S. resident paid alimony to his British ex-wife from an English bank. There, the Tax Court’s opinion could be read to indicate that alimony was somehow interest-like because it is not a fee for services: Especially instructive for our purposes is the source-of-income rule provided by statute for the payment of interest, for, like alimony, interest involves an obligation, usually to make periodic payments over a period of time, which is not incurred in exchange for property or services. Lacking these touchstones, Congress turned to the residence of the obligor—the situs of the debt—as the place where the income is produced, and thus the source of the income. The Tax Court in Container Corp. refused to read Howkins for the proposition that all types of income for which no sourcing rule is available must be sourced like interest, and thus considered whether services might provide a useful analogy: International paid Vitro to guarantee the 1991 International senior notes. These fees compensated Vitro for incurring a contingent future obligation to either pay International’s debt or make a capital contribution. Vitro was able to make this promise because it had sufficient Mexican assets — and its Mexican corporate management had a sufficient reputation for using those assets productively — to augment International’s credit and enable the long and complex series of financings we charted at the beginning of this opinion to keep going as long as it did. So we conclude that it is Vitro’s promise and its Mexican assets that produced the guaranty fees. * * * * We think that makes guaranties more analogous to services. Guaranties, like services, are produced by the obligee and so, like services, should be sourced to the location of the obligee. See secs. 861(a)(3), 862(a)(3); Hunt, 90 T.C. at 1301. We realize that we are deciding a close question, but an analogy to interest has too many shortcomings: Guaranty fees do not approximate the interest on a loan; Vitro, not International, produced the guaranty fees; and Vitro’s guaranty was not an obligation to pay immediately, but a promise to possibly perform a future act. [Footnote references omitted.] Thus, the Tax Court concluded that Vitro’s guarantee fees should be sourced like services. In that the financial assets that were considered to perform such services were located outside the United States, this resulted in the fees having a foreign source. Withholding under Search7RH1442 was, therefore, not required. COMMENTS ON THE DECISION Limited Applicability? As indicated above, the Tax Court’s conclusion that the guarantee fees at issue are unlike interest was premised in large part on its characterization of the guarantee as a contingent, secondary obligation and, moreover, one that did not necessarily increase Vitro’s liabilities or put its assets at risk, as Vitro had legal recourse against International in case the guarantee was triggered. Typically, however, modern commercial guarantees cut back on many aspects of traditional guarantees and may bear a far greater resemblance to primary obligations of a co-obligor. Where such modern guarantees are involved, the applicability of the Container Corp. decision is unclear. Accordingly, even if Container Corp. remains intact (e.g., without being legislatively overruled), multinational companies seeking to fall within that decision may wish to take a close look at the particulars of their guarantees. For example, to the extent feasible, it would be desirable to require the lender to pursue remedies against the borrower before being permitted to make any claim on the guarantee, and to give the guarantor broad subrogation rights. Tax Court’s Source Analysis Clearly, the Tax Court’s resort to the established method of sourcing by analogy was appropriate. Nevertheless, the conclusion that guarantee fees are more akin to compensation for services than to interest—or to any other type of income for which we have a sourcing rule—appears poorly reasoned. At first blush, the analogy between interest and guarantee fees appears quite on point. One is a fee for making a loan, and the other is a fee for standing ready to, in effect, make a loan in the future if certain contingencies occur; also, both are both based in large part on the creditworthiness of the borrower. The Tax Court is, of course, entirely correct in pointing out that an actual loan and a contingent obligation to make a loan are different things, but it is not like trying to compare apples and hardware stores. Indeed, the better comparison may be between apples and oranges, which, after all, are both fruits. Why the Tax Court was unwilling or unable to see some similarity is unclear. As noted above, the Tax Court saw greater merit in the services analogy, because it considered both to be “produced by the oblige,” and thought quite little of the comparison to interest. The court’s reasoning does not persuade. First, it is not clear what is meant by the assertion that guarantees, like services, are “produced by the obligee” or why this is of greater significance than the fact that guarantee fees are received from the obligor in connection with risk of the obligor’s default. Second, the court’s conclusion that guarantee fees are like compensation for services appears to be grounded almost entirely on the conclusion that such fees are unlike interest. One could argue no less persuasively (and equally arbitrarily) that guarantee fees must be similar to interest because they don’t look very much like compensation for services, as the Tax Court basically did in Howkins. Although not relied upon as a basis for its decision, it is interesting to speculate on whether the court was influenced by the fact that the IRS had previously concluded that guarantees actually constitute services. As indicated above, GCM 38499 concluded in no uncertain terms that a guarantee “is the performance of a service[,]” and thus held (conveniently) that a parent corporation guaranteeing its subsidiary’s debt for no consideration could be forced to recognize deemed income under certain provisions of the transfer pricing regulations that deal with services. Because the IRS had previously gone further by taking the view that guarantees are services, perhaps the Tax Court felt that the IRS could hardly say it is unreasonable to characterize guarantees as service-like. The Best Analogy Surprisingly, the Tax Court declined to address the one analogy that would be most on point. Although the paucity of sourcing rules in the Code is quite shocking, we do happen to have one for insurance premiums, which are sourced by reference to the location of the risk. The resemblance between a guarantee and insurance is quite striking. In each case, the insurer or guarantor undertakes the contingent obligation to reimburse the insured or lender for certain losses that may. Even if the IRS failed to make the argument, the fact that the Tax Court failed to address the analogy to insurance premiums is indeed surprising. If that analogy had been considered and deemed controlling, the fees would have had a U.S. source because International, and thus the risk being insured, was located in the United States. Other Worthy Analogies Another possible analogy for the court to consider would have been certain payments under financial instruments. For example, Search7RH988 transactions are sourced according to the residence of the taxpayer (or the relevant qualified business unit). Likewise, the source of income from a notional principal contract generally is the residence of the taxpayer. Alternatively, as noted by the Tax Court in a footnote, guarantee fees could be analogized to payments of rent or royalties, because a guarantee might be viewed as a rental of the credit rating and financial backing of the guarantor. Under either of these approaches, the result would be the same as under the Tax Court’s services approach, because Vitro’s residence, credit rating and financial backing were all situated outside the United States. These analogies seem less on point than the comparison to insurance premiums, but far more persuasive than the comparison to services. Potential Planning Opportunities If the Tax Court’s opinion in Container Corp. stands, the potential planning opportunities are significant. In particular, it may be highly beneficial for foreign-based multinationals to cause U.S. subsidiaries to pay guarantee fees to foreign affiliates in low-tax jurisdictions. If such fees are considered to have a foreign source, no U.S. withholding tax will be imposed, even if little or no tax is paid in any other country. It should be noted that, under many U.S. income tax treaties, guarantee fees paid to a treaty-based affiliate would be exempt from U.S. withholding tax. For example, Article 21(1) of the 2006 U.S. Model Income Tax Convention provides as follows: 1. Items of income beneficially owned by a resident of a Contracting State, wherever arising, not dealt with in the foregoing Articles of this Convention shall be taxable only in that State. Indeed, the Mexican treaty, which permits the other (non-residence) state to tax other income, such as guarantee fees, is unusual in this respect. Thus, it might seem that, for a foreign-based multinational group headed by a corporation organized in a country with an “Other Income” provision conforming to the 2006 U.S. Model Income Tax Convention, the favorable result in Container Corp. should not be necessary, because the applicable treaty could be invoked in any event to avoid U.S. withholding tax. That is certainly true, but this result may be less than ideal from a worldwide perspective. In order to invoke the applicable treaty, the guarantor would need to be a resident of the applicable treaty country and the beneficial owner of the income; so the guarantee fee would presumably be taxable in such country. Reducing its U.S. tax base might be nice for the guarantor, but if the cost is an increase to the foreign tax base in a country with comparable rates, that would tend to cancel out the net benefit. Obviously, it would be far more desirable to achieve the same tax benefit in the United States, but pay the guarantee fee to, for example, a Cayman Islands company, provided that this would avoid foreign tax after taking into account any CFC-type rules under the applicable non-U.S. tax laws. If the decision stands, Container Corp. would seem to make this possible; at least for traditional guarantees. An additional benefit of using guarantee payments to minimize the U.S. tax base is that the “earnings-stripping” rules would not apply. Pursuant to Search7RH163(j), a corporation’s ability to deduct certain payments of interest is subject to significant limitations; and any deduction for “disqualified interest” is generally disallowed (and carried forward to the succeeding taxable year) to the extent of the corporation’s “excess interest expense” for the year. A guarantee fee not being interest, however, the limitations of Search7RH163(j) are not relevant. Future of Container Corp. The Tax Court’s decision in Container Corp. appears to be incorrect, but in a way it serves Congress right for randomly enacting only a few source provisions and leaving it to the courts to decide what rules they like best for each scenario not governed by those provisions. Apparently, the point was well taken, as amendments to Search7RHSearch7RH861 and 862 have already been proposed to source guarantee fees in the same manner as interest. Some legislation along these lines likely will be enacted in the near future. CONCLUSION The Tax Court’s decision in Container Corp. came as a welcome surprise to many international tax practitioners, and the potential planning opportunities are very exciting. Whether any of those great plans will successfully come to fruition, however, seems doubtful.