Back to Taxes Committee publications

Bridget English
Gibson Dunn & Crutcher, London
benglish@gibsondunn.com

 

Report on a session at the 9th Annual IBA Finance and Capital Markets Tax Conference in London

Monday 20 January 2020

 

Session chair

Eric Sloan  Gibson Dunn & Crutcher, New York

 

Speakers

Mariana Eguiarte Morett  Sánchez Devanny, Mexico City

Torsten Engers  Flick Gocke Schaumburg, Frankfurt

Wiebe Dijkstra  De Brauw Blackstone Westbroek, Amsterdam

Jonathan Cooklin  Davis Polk & Wardwell, London

 

Eric Sloan explained that, while the discussion regarding the use of partnerships in cross-border transactions would be addressed on a jurisdiction-by-jurisdiction basis, there were common themes which would be addressed by the speakers, including:

• anti-hybrid rules;
• tax transparency rules; and
• anti-avoidance rules.

Mexico

Mariana Eguiarte Morett explained that Mexico has been very active in incorporating base erosion profit shifting (BEPS)-led modifications into its domestic rules.

Anti-hybrid rules

Mexico first introduced anti-hybrid rules in 2014 before the culmination of the BEPS project. Under the rules, hybrid payments to related parties which, on receipt, were subject to a preferential (ie, low) tax regime would not be deductible for Mexican tax purposes. However, the rules were somewhat limited in scope and only applied to hybrid payments of interest, royalties or ‘technical assistance’.

The rules are now being amended:

• to apply to all payments (not merely payments of interest or royalties); and
• to apply to ‘structured payments’ between unrelated parties (ie, where the payment is passed from the third-party recipient to a related party of the payer).

Mexican tax authorities are expected to use the exchange of information provisions in their tax treaties to audit these kinds of transactions.

Transparency rules

Recently published proposals indicate tax transparency rules in Mexico will soon change.  

Based on certain regulations, it had been interpreted that Mexico recognises the tax transparency of ‘legal figures’ (ie, entities which lack legal personality and tax personality). Canadian limited partnerships are a typical feature of structures investing in Mexico because the Mexican tax authorities have previously confirmed that such entities would be treated as transparent for Mexican tax purposes. Other entities could, in principle, benefit from the same treatment if they met the relevant requirements but, in the absence of an equivalent confirmation from the Mexican tax authorities, taxpayers would not enjoy the same level of certainty.

Pursuant to the above-mentioned recently published proposals, certain vehicles currently treated as transparent – such as Canadian limited partnerships – will no longer generally qualify as such as of 2021. They will instead be treated as opaque for Mexican tax purposes.

The amendments to the transparency rules was the subject of significant discussion within the private equity industry, which negotiated with the Mexican tax authorities to obtain certain concessions (a so-called ‘tax benefit’): limited partnerships in private equity structures receiving certain kinds of income (dividends, interest, capital gains, real estate income) will be recognised as transparent.

However, the concession comes with considerable administrative burdens. The new rules disregarding transparency were introduced because Mexican tax authorities weren’t comfortable that they had sufficient knowledge of the ultimate beneficial owners behind the transparent entities. Accordingly, transparency will only be recognised if:

• the vehicle is formed in, and its investors are resident in, a jurisdiction with an exchange of information agreement with Mexico; and
• significant information is filed with the Mexican tax authorities.

As per current proposals, the concession would only allow transparency to be recognised at one level. This is because transparency will only be recognised if the limited partners are the beneficial owners of the income. In practice, common tiered structures (eg, partnerships in which ‘feeder’ funds are partners) would not benefit. The private equity industry continues to lobby for a wider concession which would apply to structures involving multiple partnerships and to emphasise that such structures are often needed for non-tax reasons.

The proposed changes involve Mexican domestic law only. However, Mexico is party to a number of double tax treaties which may alter its application. For example, Mexico has a mutual agreement procedure in place with the United States which recognises the tax transparency of certain vehicles regardless of the position under domestic rules. The above changes would not apply to investors benefiting from the US-Mexico double tax treaty. Where there is a mixed fund featuring US and non-US investors, the benefit of the mutual agreement procedure would be recognised in proportion to the portion of US investors in the fund. Depending on the identity of the other investors, other double tax treaties might also override the changes in Mexico’s domestic rules.

Germany

Anti-hybrid rules

Torsten Engers explained that Germany is in breach of its obligations under the second Anti-Tax Avoidance Directive (ATAD II) as it has not yet implemented anti-hybrid rules. The draft rules that have been published merely transpose the text of the Directive. Accordingly, there is some uncertainty regarding the approach the German tax authorities will take in enforcing the rules once they take effect.

Transparency rules

Engers provided an overview of German transparency rules as they apply to partnerships, noting certain novel features of the German regime.

Allocation of income

In principle, partnerships are transparent for German tax purposes. However, German tax rules governing the allocation of income for tax purposes have such a broad application that taxable income may nevertheless be allocated to partnerships. Engers used the example in Figure 1 to illustrate this.

Figure 1

 

Looking first to the payment of interest by the German partnership to the first Dutch private company (BV1), ‘transparency’ would be achieved as the interest income would be allocated to, and would also constitute a deductible expense of, the partnership. Where interest articles in treaties allocate taxing rights only to the recipient jurisdiction, the German tax authorities have historically argued that the income was business income (rather than interest income), which Germany was free to tax. However, domestic tax courts disagreed with that position. In such circumstances, it is necessary for the Germany tax authorities to apply a treaty override; the interest income will be subject to German tax, but a tax credit will be allowed against taxes paid in the counterparty jurisdiction (in Figure 1, Dutch tax paid by BV1).

Turning to the loan from BV2 to BV1, the interest expense of BV1 thereunder would also be taken into account to determine the income of the Kommanditgesellschaft (ie, limited partnership business entity) KG for German tax purposes. Though not an expense of the KG, it would be allocated to the KG for German tax purposes. These rules were historically used to a taxpayer’s advantage; from a Dutch point of view, a fiscal unity could be formed between BV1 and BV2 so that the interest income was not recognised for Dutch purposes while giving rise to a deduction for German tax purposes. However, rules have been implemented in both the Netherlands and Germany to ensure such planning is no longer effective.

Engers then explained the application of German tax rules to outbound structures involving partnerships, in which a German taxpayer holds shares in a foreign corporation via a foreign partnership (see Figure 2):

Figure 2

 

The German taxpayer’s intention here is that the dividend income would be allocated to the Dutch partnership, which qualifies as a permanent establishment of the German taxpayer. In such circumstances, there should be no corporate income taxation in Germany. However, there are a number of German court decisions (going back seven or eight years) which conclude that the mere fact that legal title to the shares is held by the Dutch partnership does not automatically mean that income from the shares must, for German tax purposes, be allocated to the Dutch partnership.

Under the relevant German rules, shares owned by a taxpayer are allocated to the taxpayer’s jurisdiction of residence unless there is a ‘functional relationship’ between the shares and a permanent establishment of the taxpayer. For a functional relationship to exist, there usually must be a kind of business relationship between the partnership and the company in which the shares are held (eg, where one acts as a distributing entity in the group and the other acts as a production entity, or where the partnership has management functions toward the company).

In this example, in the absence of a functional relationship, the dividend income would be allocated to the German taxpayer. That could produce a harsh result. For example, as the Dutch partnership’s shareholding in BV1 falls below the ten per cent threshold necessary to qualify for the German dividend exemption, the dividend would be fully subject to German corporate income tax.

Application of treaties

Engers noted a second ‘peculiarity’ in the application of German tax rules to partnerships involving the interpretation of the dividend article in the European Union Parent-Subsidiary Directive. Subject to certain conditions, the Directive is intended to provide a tax exemption for dividends from direct participations where both the payer and the recipient are resident in the EU. For German tax purposes, a foreign investor would lack a direct participation in the Germany entity if a partnership sits between the German entity distributing the dividend and the foreign investor. The investor would not be able to rely on the Directive to avail itself of favourable tax treatment.

The same approach is taken in applying the interest article in German double tax treaties. However, in the last month, the lower German courts have ruled against this approach of denying the benefit of the treaty where a civil law partnership has been interposed. An appeal is ongoing.

Finally, Engers asked the other panellists about the application of tax treaties in their jurisdictions in circumstances where a partnership sits between the investor and the entity paying the dividends. Wiebe Dijkstra noted that such an investor would be able to rely on the treaty between the Netherlands and their jurisdiction of residence. However, Dutch entity classification rules are quite strict and, in the absence of particular structuring, a partnership would typically be opaque.

Sloan explained that the application of US tax rules for dividends received through a partnership are unclear.

The Netherlands

Anti-hybrid rules

Dijkstra explained that, following BEPS Action 2, the Netherlands had implemented anti-hybrid rules, in keeping with its obligations under ATAD II (the Netherlands previously had only one minor rule of this kind, further to the ‘anti-hybrid instruments’ rule included in the EU Parent-Subsidiary Directive). In his opinion, the rules, which took effect on 1 January 2020, are a ‘trap for the unwary’ as the Netherlands has implemented a ‘gold-plated’ standard. The only positive exception (from a taxpayer perspective) is the lenient approach taken toward special regimes.

The following aspects of the Dutch regime are noteworthy:

• Income which has been subject to a controlled foreign corporation (CFC) income inclusion does not easily qualify as income which has been ‘taxed’ for the purposes of determining whether there is a hybrid mismatch. The rules raise the prospect that income which has been taxed under the US foreign-derived intangible income (FDII) rules would qualify as ‘taxed’ for this purpose, but income taxed under the US global intangible low-taxed income (GILTI) rules would not. In Dijkstra’s view, the logic behind the difference in treatment is unclear.
• The Dutch regime incorporates strict documentation requirements. If a taxpayer cannot positively show that a payment does not give rise to a mismatch, then there is a reversal of the burden of proof.

The application of the new rules was illustrated using the example in Figure 3.

Figure 3

 

In Figure 3, a Dutch BV is wholly owned by a US private real estate investment trust (REIT, which is exempt from tax if it makes sufficient distributions). The Dutch BV is checked as transparent from a US perspective, and owns shares in a Canadian real estate company, which is also checked as transparent for US purposes. The Dutch BV has taken out two loans:

• one from its indirect parent, a US partnership which is held by as US-listed REIT and
• one from a third-party bank.

The BV receives 150 of interest from the Canadian real estate company. 50 of that interest is used to pay interest to the third-party bank, and 100 is used to pay interest to its indirect parent, the US partnership.

Bank loan

The starting point is that there is a double deduction because the bank interest is deductible at the level of both the Dutch BV and the US REIT. It is therefore necessary to consider whether the exemption for dual inclusion income applies.

Similar examples in technical explanations from the Dutch Secretary of Finance indicate that the 150 of income will not be considered dual inclusion income. This is because (as both are checked open) the US would disregard the interest paid by the Canadian real estate company to the BV. Although the US does tax the operating income of the Canadian real estate company, the Dutch tax authorities consider that to be a different form of income.

Accordingly, the exclusion cannot apply, and a deduction is denied for the 50 of interest paid by the Dutch BV to the bank. In Dijkstra’s view, the Dutch tax authorities’ approach here is representative of a gradual, but noticeable, shift away from their previously facilitatory and pragmatic attitude to commercial transactions.

Loan from US partnership

Again, the exception for dual inclusion income will not apply (for the same reasons). An additional issue may also arise. The US partnership income is included at the level of the listed US REIT on a net, rather than gross, basis (ie, the US listed REIT is only subject to tax on the amount equal to the partnership’s income minus the partnership’s deductible expenses). Although the point is not expressly covered in the publications of the Dutch Secretary of Finance, interest paid to the US partnership may not qualify as ‘included’ at the level of the partnership.

Dutch advisers had hoped to be able to resolve these issues with the Dutch tax authorities, but they were unsuccessful.

Anti-avoidance rules

Dijkstra also provided a brief overview of new Dutch anti-avoidance proposals.

The Dutch government has decided to introduce a new withholding tax on interest and royalties (withholding tax on dividends already applies, subject to certain exclusions.) The measures are not intended to generate revenue but instead to deter the use of the Netherlands in conduit structures (eg, group financing and licencing structures involving certain low-tax jurisdictions). Withholding tax will be levied at the current corporate income tax rate on intra-group payments to certain blacklisted jurisdictions (eg, Bermuda or the British Virgin Islands). In Dijkstra’s experience, most groups are already careful to give a wide berth to such structures in light of the heavy penalties.

The new rules are also capable of applying in circumstances which do not involve blacklisted jurisdictions. There are particular risks for partnerships.

This was illustrated using the example of a typical hybrid structure involving a foreign partnership which is treated as transparent from the perspective of its investor partners, but is opaque from a Dutch tax perspective (as foreign partnerships generally are). The default position is that the new withholding tax will apply, unless certain ‘rebuttals’ can be established.

For a typical hybrid structure, such ‘rebuttal’ can be successfully made if every partner in the partnership which is not an individual and which has a qualifying (ie, controlling) interest:

• is not resident in a low tax jurisdiction (ie, a jurisdiction with a corporate income tax (CIT) rate below nine per cent or an EU blacklisted jurisdiction); and
• is the beneficial owner of the interest and royalties under the rules of their jurisdiction of residence.

A couple of issues are of particular note. In principle, it will be necessary to determine, on an investor by investor basis, who holds a controlling interest. For this purpose, the interests of investors who ‘cooperate’ must be aggregated. Therefore, it is important to understand when investors will be considered to ‘cooperate’. It could be argued that the existence of a partnership would itself be sufficient, although that is too aggressive an interpretation in Dijkstra’s view.

Additionally, a ‘rebuttal’ cannot be applied proportionately; if any investor does not meet the requirements, the rebuttal will not be made and the withholding tax will, subject to the application of any double tax treaty, apply in full (US investors may still rely on the transparency rules in the US/Netherlands double tax treaty, for example.)

A second example was discussed, involving a ‘reverse hybrid’ structure (ie, a partnership which is treated as opaque from the perspective of the investors, but as transparent for Dutch tax purposes).

Again, the default is that withholding tax would apply but, a ‘rebuttal’ can be established to exempt the dividend if the investors are:

• treated as the beneficial owners of the dividend under the rules of their jurisdictions of residence; and
• not resident in blacklisted jurisdictions.

If any investor fails to meet these conditions, a proportionate approach can apply (so that withholding tax would only apply to the portion of the dividend ultimately owned by that investor). This is because the Dutch tax authorities accept the partnership as transparent and so apply the rules on an investor-by-investor basis.

United Kingdom

Basic principles of UK taxation of partnerships

Jonathan Cooklin highlighted some general features of UK partnership taxation:

Merely describing an entity as a partnership does not mean it will constitute one for UK tax purposes. A partnership is the relationship which exists between persons carrying on a business in common, with a view of profit. Labels are not determinative; it is necessary to consider the legal rights, obligations and attributes. UK tax authorities may, for example, query whether an alleged ‘partnership’ is:

• carrying on a business at all (as opposed to simply being a passive entity); or
• if so, whether it is doing so with a view of profit.

Although quite different, both a Scottish limited partnership and a Delaware limited partnership generally qualify as partnerships for UK purposes.

In addition, no uniform code governs the treatment of partnerships under UK tax law. Rather, the tax treatment is highly context-specific. Partnerships are not always treated as transparent (although for the purposes of some rules they are) and entities treated as transparent need not necessarily be partnerships. Whether an entity is a partnership may, in some instances, be irrelevant to its tax treatment, which could depend on other factors such as whether it has legal personality or is a body corporate. 

Cooklin gave a recent example of a circumstance where the existence of a partnership had thrown up some thorny issues. The (simplified) example involved a debt restructuring of a group in financial difficulty (see Figure 4).

 

Figure 4

 

A UK company and a US company were partners in a non-UK limited liability partnership (LLP) which owned a worldwide group. The partnership had borrowed funds from a bank and was paying interest. UK rules governing the taxation of loan relationships treat the loan (in form, a debt obligation of the LLP) as a loan of the US partner and a loan of the UK partner. That is relevant for a number of reasons and raises a number of issues.

First, the UK has purpose-based anti-avoidance rules. If the main purpose of a transaction is to secure a tax deduction, then that deduction may be denied. Here, the question arises as to whether it is the purpose of the LLP or the purpose of the UK partner (ie, the deemed debtor), which should be considered?

Second, in the context of a debt restructuring, special favourable rules apply to insolvent debtors or on the conversion of debt into equity. Those rules strictly only apply to loans of a company. Does the above deeming rule allow the debt to be considered a debt of the UK partner for this purpose, notwithstanding that it is, in form, a debt of the LLP?

Third, the LLP is treated as transparent for the purposes of the UK loan relationship rules. However, the withholding tax position applying to interest under the loan is governed by a different regime, which applies withholding on UK source interest. In Cooklin’s view, it is relatively clear that a partnership is not treated as transparent for the purposes of those rules. Interest paid by a non-UK LLP (which constitutes a body corporate or has separate legal personality) should not, in principle, be UK source interest and so should not attract UK withholding tax.  

Finally, and although not relevant in this example, tracing through different partnerships in applying the deeming rule is difficult, as the relevant provisions do not readily accommodate tiered structures. Nevertheless, it appears that is how HM Revenue & Customs (HMRC) intend the rules to operate.

Anti-hybrid rules

Cooklin noted that, as with the Netherlands, the UK had adopted a gold-plated hybrid-mismatches regime. The rules have proven particularly problematic when advising US multinationals, given that ‘checking the box’ and the resultant hybridity is part of the DNA of the US tax regime. Hybridity is not abusive in the US – it is just part of its system. In Cooklin’s experience, HMRC are unmoved by this point and have been unable to offer any practical suggestion as to how the difficulty should be overcome.

United States

Finally, Sloan provided an overview of some impending changes to the US tax treatment of partnerships.

CFC rules

The Internal Revenue Service (IRS) intends to amend how the US CFC rules will apply to partnerships. In Sloan’s view, the proposals are both significant and sensible. Currently, if a US partnership owns a non-US company, the latter would be a US CFC (subject to US Subpart F rules). This is the case even if, on a look-through basis, the partnership is not controlled by a US taxpayer.

The new proposals would still treat a foreign corporation, owned by a US LP, as a US CFC. However, partners in the US LP would only be subject to income inclusion (under Subpart F rules) if the partners own ten per cent or more of the partnership. The regulations are in proposed form and some further changes can be expected.

Interest deductibility

New limits are to be introduced on interest deductibility under section 163(j) (a repurposed Code section). Interest will be limited to 30 per cent of earnings before interest, tax, depreciation and amortisation (EBITDA) until 2022, and will then drop to 30 per cent of EBIT.

In what Sloan views as a conceptual error (also made by the German government in applying equivalent rules), the legislation adopted by Congress requires the test to be applied at the level of the partnership, rather than the partners. It has taken much effort from the IRS (and complicated rules involving an 11-step process) to attempt to manage that mistake. Final regulations are expected in the next month.

Certain features of the proposals are particularly noteworthy. First, the limitations will not apply if debt is converted to equity. This is the case even for preferred equity, as its form is typically respected. While there are some rules in the proposed regulations which might hinder that outcome, Sloan expressed his view that such rules are almost certainly invalid and there are indications they will not feature in the final regulations.

In addition, the fact that the rules are applied at partnership level may create some flexibility to reorganise groups to obtain a more favourable outcome. For example, where one entity has a lot of EBITDA and little interest expense, and another has little EBITDA but a lot of interest expenses, it might be possible to combine the two in partnership to reduce the extent of the interest restriction.

BEAT minimum tax

The base erosion anti-abuse tax (BEAT) rules contain special provisions for partnerships, including some clear ‘no-fly zones’. Particular caution is needed if a partnership with both US and non-US related persons obtains depreciable property.

 

Back to Taxes Committee publications