The Multilateral Instrument
On June 7, 2017, the formal signing ceremony of the Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting (the “Multilateral Instrument” or “MLI”) took place. Sixty-eight jurisdictions have signed the MLI, with another nine jurisdictions signing a letter indicating their intent to sign the MLI.
The MLI is one of the outcomes of the OECD/G20 Project to tackle Base Erosion and Profit Shifting (the “BEPS Project”). Specifically, Action 15 of the BEPS Project focused on the development of an MLI to enable countries to efficiently modify their bilateral tax treaties to implement certain BEPS measures. What could otherwise take decades to accomplish (i.e., separately renegotiating 3,000 bilateral income tax treaties to accommodate BEPS), the MLI does with a single multilateral instrument.
Although the United States has not signed the MLI, the MLI nevertheless may dramatically impact many inbound structures utilized by foreign investors in the United States, as well as many outbound structures used by U.S.-based multinationals. For example, the possible inclusion of an anti-triangular provision from the MLI in a particular treaty could prevent treaty benefits from being available (and thus result in increased withholding tax) where payments are made to a third-country permanent establishment (PE) and that PE is not subject to a sufficient rate of tax. Many existing structures also could be adversely affected by the inclusion of certain new elimination of double taxation provisions contained in the MLI. Under these provisions, an exemption may be denied in the home country for income that “may be taxed” in the treaty partner country, with a credit being provided instead for the tax paid on such income in the other country (which may often be zero or a de minimis amount).
Despite these possible changes, as further discussed below, a number of interesting planning opportunities appear to remain.
Common Planning Techniques Utilizing Foreign-to-Foreign Treaties
To illustrate the potential impact of the MLI on many commonly used structures, consider the following examples. Assume a U.S.-parented multinational has subsidiary operations in several foreign jurisdictions in Latin America and Europe, including Mexico and Poland.
In an effort to minimize corporate income tax in Mexico, the Polish subsidiary sets up a low-taxed finance branch in Switzerland that constitutes a PE for treaty purposes. The branch then makes loans to the Mexican subsidiary in exchange for a promissory note bearing stated interest.
Ignoring the MLI, the tax implications of the above structure generally would be as follows. The interest payments on the note should be deductible for Mexican corporate tax purposes, and under domestic law should be subject to a 35% withholding tax in Mexico. Article 11 of the Mexico-Poland income tax treaty, however, reduces this 35% Mexican withholding tax rate to 15%. These interest payments should be exempt from tax in Poland, because Article 23(1) of the Poland-Switzerland income tax treaty prohibits Poland from imposing tax on income that “may be taxed” in Switzerland. And in this example, the income is attributable to a Swiss PE, and therefore “may be taxed” in Switzerland pursuant to Article 7 of the Poland-Switzerland treaty (notwithstanding that any Swiss taxes would be imposed at a nominal rate). This structure thus allows the group to reduce its global tax burden by allocating income from a high-tax jurisdiction (Mexico) to a low-tax jurisdiction (Switzerland), and relying on the special provisions of a treaty to avoid triggering tax on that income in other jurisdictions (i.e., Poland).
Implications of MLI Article 10
The MLI, which Mexico, Poland, and Switzerland all have signed, may impact the above structure as follows. Under Article 10 of the MLI (entitled, “Anti-abuse Rule for Permanent Establishments Situated in Third Jurisdictions”), any income paid from Mexico to Poland and allocated to Switzerland will not be eligible for treaty benefits in Mexico if the tax rate in Switzerland is less than 60% of the rate that would have applied in Poland. Under the above facts, the nominal Swiss tax imposed by a Swiss finance branch will be less than 60% of the 19% corporate tax rate that would have applied in Poland. Therefore, pursuant to Article 10 of the MLI, treaty benefits should be denied. This would result in a 35% withholding tax being applied to the interest payments in Mexico, essentially rendering pointless this particular structuring.
It may be possible to avoid the above adverse result under the MLI and still achieve a similar outcome by using a Maltese finance branch in place of the Swiss branch used in the above example. As above, the interest payments on the note should be deductible for Mexican corporate tax purposes, and also should be exempt from tax in Poland (because Article 23(1) of the Poland-Malta income tax treaty, like the same provision of the Poland-Switzerland treaty, prohibits Poland from imposing tax on income that “may be taxed” in Malta. In this case, the income is attributable to a PE in Malta, and therefore may be taxed in Malta pursuant to Article 7 of the treaty).
While Malta also has signed the MLI, a question arises as to how to measure the tax rate in Malta for purposes of the 60% comparison test set out under Article 10 of the MLI (hereafter, the “rate comparison test”). Malta has a rather unusual domestic tax law system, under which the Maltese PE initially would be subjected to a 35% tax in Malta on its income, but when a distribution of profits was later made from the branch to its foreign owner, in most cases 6/7 of the Maltese tax would be refunded to the recipient (leaving only a 5% effective tax burden in Malta). It is not clear which tax rate, the 35% rate or the 5% rate, would be considered applicable in Malta for purposes of the rate comparison test. While at first glance, it may seem that the 5% rate should be controlling, this is far from clear, given that the 35% rate is relevant for other purposes within the international tax context (for example, the CFC rules of many countries).
An analogous issue arises in the case of the use of an Estonian branch of a corporation resident in certain countries. Assume the above example involved an Estonian branch of a Hungarian company, rather than a Maltese branch of a Poland company. Estonia also has a unique tax system, which does not impose any corporate income tax on a resident company as the income is earned, but instead imposes a 20% corporate income tax on that income at the time of distribution. This rule also applies to an Estonian branch of a non-resident company. Thus, for example, if an Estonian branch of a Hungarian company makes loans to the Mexican subsidiary discussed above, interest payments on the loan should be eligible for reduced withholding tax rates under the Hungary-Mexico income tax treaty, should be exempt from tax in Hungary (since it is allocated to a PE in Estonia), and should not be subject to current tax in Estonia (based on the unique Estonian corporate tax regime). The question arises under this scenario, as under the Poland-Malta scenario immediately above, as to which tax rate in Estonia is controlling for purposes of the rate comparison test – the 0% rate (which clearly would fail the rate comparison test, resulting in treaty benefits being unavailable) or the 20% rate (which would satisfy the rate comparison test).
Further Implications of MLI Article 5 on the Above Structures
The above discussion of Article 10 of the MLI focused on the MLI’s disallowance of reduced withholding tax rates otherwise available under a treaty. Article 5, Option C of the MLI (where this option is selected by a relevant MLI jurisdiction) may further impact these structures by eliminating an exemption that would otherwise apply to income earned through a PE.
Option C of Article 5 of the MLI provides that, where a resident of a treaty jurisdiction derives income which may be taxed in the other contracting state, the first-mentioned jurisdiction will grant a foreign tax credit (rather than a complete exemption) on such income. Therefore, based on the above examples, if Poland and Hungary chose Option C of Article 5, those countries would not exempt the income of the finance branch from tax, but rather would simply provide a foreign tax credit for the taxes paid in Switzerland, Malta, or Estonia, respectively. In any of these cases, the tax paid by the branch is nominal or nil, such that this consequence would be quite significant and would again essentially defeat the tax-planning objectives of the structure.
Double Non-taxation Opportunities Not Impacted by the MLI
While it is clear that one of the principal goals of the MLI is to curb abusive instances of double non-taxation, the MLI does not eliminate all such possibilities. Treaties, when combined with domestic law, may still permit double non-taxation in cases where a treaty partner country “may” tax income, but under its domestic law (rather than pursuant to a treaty provision) does not exercise its discretion to do so and either the country granting the exemption is not a party to the MLI, or the MLI jurisdiction in question does not select Option C of Article 5 of the MLI.
For example, assume a U.S. parent corporation owns a Dutch subsidiary and a Venezuelan subsidiary. As of the date of drafting this post, Venezuela has not signed the MLI. Assume the Venezuelan subsidiary licenses certain IP to the Dutch subsidiary in exchange for royalty payments, or makes a loan to the Dutch subsidiary and receives interest payments on the relevant debt. Article 24(5) of the Netherlands-Venezuela treaty provides that, “[w]hen a resident of Venezuela receives income which according to the provisions of this Convention may be taxed in the Netherlands, this income shall be exempt from Venezuelan tax.” Articles 11 and 12 of the treaty further provide that the Netherlands “may” tax interest and royalties sourced in the Netherlands. Accordingly, Venezuela does not have the right to impose tax on these payments. However, Dutch domestic law does not impose any withholding tax on such outbound interest or royalties. Accordingly, neither Venezuela nor the Netherlands imposes any tax on the royalty or interest payments. This instance of double non-taxation will not be affected by the MLI, given that Venezuela is not a party to the MLI. And even if Venezuela does sign the MLI, this structure does not appear to be affected by Article 10 of the MLI, and should only be affected by Article 5 in the event Venezuela chooses Option C.
Another example of this potential double non-taxation benefit, in this case with respect to dividends, may arise under the Brazil-Spain treaty. Brazil also is not currently a party to the MLI. Article 23(4) of the Brazil-Spain treaty provides that “[w]here a resident of Brazil derives dividends which, in accordance with the provisions of this Convention may be taxed in Spain, Brazil shall exempt such dividends from tax.” Article 10 of the treaty further provides that Spain “may” tax Spanish-sourced dividends, which means Brazil has no right to tax these amounts. However, pursuant to local Spanish law, dividends paid by an ETVE (a type of Spanish holding company) are not subject to withholding tax. As a result, neither Spain nor Brazil would impose tax on dividends paid by a Spanish ETVE to its Brazilian parent company. As in the preceding example dealing with Venezuela, this instance of double non-taxation would not be affected by the MLI, given that Brazil is not a party to the MLI. And even if Brazil does sign the MLI, this exemption should only be affected by Article 5 of the MLI in the event Brazil chooses Option C.
As illustrated above, the MLI goes a long way toward curbing the types of abusive structures it was designed to eliminate. Given that not all countries have signed the MLI or consistently adopted similar treaty provisions on a bilateral basis, however, many such structures appear to remain viable.