On February 17, 2010, the U.S. Tax Court released its decision in Container Corporation v. Commissioner of Internal Revenue,1 which clarifies, if not changes, prior law regarding the source of fees paid for financial guarantees for U.S. federal income tax purposes. The decision, if not reversed on appeal, could be seen as clarifying that U.S. withholding tax is not typically required on the cross-border payment of such fees, regardless of whether a tax treaty applies to such income. While the recent Fifth Protocol to the Canada-U.S. tax treaty (the Treaty) greatly simplifies the withholding analysis for guarantee fees (generally precluding the application of the U.S. withholding tax regime to most such intra-group cross-border fees), this case provides the possibility of additional relief for Canadian companies receiving payments of cross-border guarantee fees that are otherwise ineligible for Treaty benefits, such as guarantee fees derived through a hybrid entity or paid to a Canadian parent corporation that does not satisfy the limitation on benefits provisions of the Treaty.

Like most countries, the United States imposes non-resident withholding taxes. The United States imposes such taxes (at a rate of 30%, unless reduced by relevant treaty) on U.S.-source income paid to non-U.S. persons that is not connected with a U.S. trade or business. Although the U.S. Internal Revenue Code (the Code) provides rules for classifying income as U.S. or non-U.S. source, the rules are not exhaustive. As a result, U.S. courts have been forced to rely on comparison and analogy in determining the source of those categories of income that are not specifically addressed by the statute,2 including the source of guarantee fees.3 The tax characterization of fees for guarantees or other forms of credit support between U.S. and Canadian affiliates has, accordingly, been an issue in cross-border tax planning for some time.

In Bank of America v. United States,4 the U.S. Tax Court struggled with the source of acceptance and confirmation commissions paid by foreign banks to Bank of America (BofA) under letters and lines of credit. In that somewhat limited factual scenario, the court concluded that the fees should be sourced by analogy to interest, reasoning that they were essentially compensation for the substitution and use of BofA’s credit, as though BofA had assumed the credit risk and made direct loans to the foreign banks requesting the acceptances and confirmations.5 Accordingly, because interest is generally sourced under the Code to the residence of the borrower, the Bank of America case suggested that guarantee fees, which could similarly be thought of as compensation for the use of the guarantor’s credit, should similarly be sourced to the borrower’s jurisdiction (therefore causing the acceptance and confirmation fees in Bank of America to be non-U.S. source). Stated differently, the Bank of America case suggested that fees paid by a U.S. subsidiary to its Canadian parent as compensation for the Canadian parent’s guarantee of the U.S. subsidiary’s debt should be U.S.-source and therefore should be subject to U.S. withholding tax (subject to potential reduction under the Canada-U.S. treaty).

In contrast, the Tax Court in Container instead views guarantee fees as analogous to services, which are generally sourced under the Code to the place where the services are performed.6 In reaching this conclusion, the court reasoned that the guarantee fees being paid by a U.S. subsidiary to compensate its Mexican parent for its guarantee of the subsidiary’s debt did not flow from the substitution of the guarantor’s credit for that of the borrower (as in Bank of America). Instead, the court reasoned that, because the guarantor was only liable in the event that the borrower defaulted (and not primarily liable as under a letter of credit), the guarantee fee was more akin to a payment for the guarantor’s promise to possibly provide services to the borrower in the future than interest on a loan being made by the guarantor to the borrower. Accordingly, the court concluded that the guarantee fee should be sourced to the location of the guarantor. Under this analysis, a guarantee fee paid by a U.S. entity to its guaranteeing Canadian parent would be treated as Canadian-source income, and therefore would not be subject to U.S. withholding tax regardless of the application of the Treaty.

As alluded to above, the recently effective Fifth Protocol to the Treaty provides that guarantee fees derived by a resident of one country are taxable only in the country of residence (unless such fees are attributable to a permanent establishment in the other country).7 Accordingly, the payment of guarantee fees to Canadian affiliates on most intra-group guarantees are no longer subject to U.S. withholding tax, given that such groups are not ordinarily engaged in the trade or business of providing guarantees through a U.S. permanent establishment. However, because the availability of Treaty benefits may be restricted under the Treaty’s limitation on benefits provisions,8 and because the new “anti-hybrid” provisions of the Fifth Protocol9 have made Treaty benefits unavailable, in certain cases, to guarantee fees derived through or paid by hybrid entities (which are frequently utilized by North-American cross-border groups to accommodate their cross-border financing objectives), the Treaty does not provide complete U.S. withholding tax relief for all guarantee fees. Unless reversed on appeal, Container could fill the gap by, as discussed above, sourcing such fees to the Canadian guarantor’s jurisdiction, thereby removing such fees from the gauntlet of the U.S. withholding tax regime.

The sourcing of guarantee fees under Container, however, could have adverse foreign tax credit implications in the reverse case where a U.S. parent corporation guarantees the debt of its Canadian subsidiary. Generally, U.S. corporations paying foreign taxes prefer non-U.S. source income, which raises the foreign tax credit limitation for utilizing creditable foreign income taxes. The Container case, however, would cause such guarantee fees to be U.S. source. This would create difficulties for payments of guarantee fees by a Canadian subsidiary to its U.S. parent in circumstances where Treaty benefits are unavailable. In such case, Canada would generally impose withholding tax on the guarantee fee because the fee is paid by a Canadian resident to a related non-resident person, but the U.S. parent’s ability to obtain a foreign tax credit in respect of such Canadian withholding taxes may be eliminated, or at least materially impaired, by virtue of the fact that the guarantee fee income is treated as U.S.-source income for U.S. federal income tax purposes. In other words, the result is a classic source mismatch, with Canada effectively treating the fee as arising in Canada and the United States treating it as U.S. source. Because of the unavailability of Treaty benefits in the example, there is no overarching bilateral rule to reconcile the sourcing inconsistency.10 Even in cases where Treaty benefits are available and, accordingly no Canadian withholding tax is exigible on the payment of the guarantee fee, the treatment of such amounts as U.S.-source income may nevertheless disrupt prior U.S. foreign tax credit planning put in place by the U.S. parent.

Interestingly, neither the Bank of America court nor the Container court chose to source the relevant guarantee fee by analogy to insurance underwriting income, which is generally sourced to the location of the insured risk. Under such a characterization, a guarantee fee would presumably be sourced to the residence of the borrower, whose failure to pay the debt gives rise to the guarantor’s potential liability under the guarantee. In addition, Container, in distinguishing Bank of America, relies heavily on the fact that the guarantor was only liable in the event of the borrower’s default (unlike BofA, which was primarily liable). Virtually equivalent economic results to a guarantee arrangement, however, could arguably be achieved (particularly in the context of a parent guarantor) by having the guarantor borrow directly from the lender (at a lower interest rate than otherwise available to the borrower, absent the guarantee) and then by having the guarantor on-lend the funds to the borrower (at the higher rate available to the borrower, absent the guarantee). In such a transaction, the guarantor would only be required to use its own funds to repay the lender if the borrower defaulted on its payment obligation to the guarantor. Accordingly, from an economic perspective, the guarantor could still be thought of as being secondarily (and not primarily) liable, but the payment of the guarantee fee would effectively have been replaced by the payment of actual interest (equal to the interest-rate spread between the guarantor’s loan to the borrower and the lender’s loan to the guarantor). Given these possible alternative characterizations, each seemingly with merit, there is some speculation that Container could be overturned on appeal. Though reliance on Container without thoughtful analysis should be cautioned, Container is currently good law and could offer planning opportunities and tax savings in the context of certain credit support transactions, particularly where Treaty benefits are unavailable.