As previously reported on this Blog the EU Anti-Tax Avoidance Package was presented on 28th January 2016. The package included:

–          Recommendation on Tax Treaties

–          Amended Directive on Mandatory Exchange of Information (CbC)

–          External Strategy for Effective Taxation

–          Anti-Tax-Avoidance Directive (ATAD)

The overall policy objective of the package is:

–          Effective taxation: Ensuring tax is paid where the value is created

–          Transparency: Ensuring effective access to tax information

–          Addressing the risk of double taxation

The ATA-Package is the joint European Union’s coordinated answer to BEPS, but is also partially a carve-out of the ATA rules contained in the 2011 CCCTB draft directive. The ATA package also serves to ensure EU-law conformity of ATA-rules. Moreover, it is stated that the package ensures the creation of a better/fairer business environment. Following intense negotiations the Council finally adopted the Package on 12 July 2016. Compared to the draft version, the final version of the ATAD contains significant changes. Implementation of the ATAD generally has to be carried out no later than 1st January 2019, with deviations regarding exit tax and interest limitation rules.

The purpose of this blog is to present some first comments to the final ATAD.

As a general remark it is yet too early to assess the final impact on the 28(27) EU MS tax systems (and no impact assessment is presented by the European Commission).

The ATAD is a minimum directive (de minimis approach), combined with multiple options for the MS. This means that MS are obliged to ensure at least the level of protection as described in the directive. However, MS cannot offer less restrictive rules. Consequently, MS are allowed to apply more restrictive rules (article 3). The ATAD contains only general provisions which leave implementation for MS in a way that fits best with their corporate tax systems.

Thorough analysis has to determine whether and which changes are needed to domestic legislation. In our view more changes may be the end result in comparison to the early comments presented to the ATAD. Some MS seemingly expect to carry out as little changes as at all possible. Having said this we sincerely believe, that the MS cannot rely on article 3 of the ATAD with the intention of doing nothing – a concrete case-by-case assessment has to be made to determine whether correct implementation is carried out. These issues should be policed by the Commission.

One question arising is whether correct implementation includes the whole rule in the ATAD or details of the rules in the ATAD. It should not be sufficient to have domestic rules of the same nature if they do not at least match the level of ATA protection as presented in the ATAD. The problem is that such an assessment often depends on the facts and circumstances of concrete cases.

In terms of the subjective scope, the ATAD is applicable to all taxpayers subject to corporate tax (entities and PEs). Consequently, the ATAD includes more taxable entities than the current EU company tax directives, including PEs of third county entities. There are still variations between EU MS in terms of which entities are subject to corporate tax and moreover, certain specific vehicles are in principle subject to tax, although objectively exempt from corporate income tax. The ATAD contains no guidance regarding this scenario.

The final ATAD now contains five anti- tax avoidance rules, after the controversial switch-over clause was abandoned in the final version. The five remaining rules are interest limitation rules, rules on exit taxation, a GAAR, CFC rules, and anti- hybrid rules, which we will comment on in the following.

Interest limitation rule

The ATAD follows the international trend towards EBITDA-based rules as a means of preventing leveraged financing. The ATAD contains a rule allowing deduction of net borrowing costs up to 30% of EBITDA (defined for tax purposes). The provision is general in its scope and does not make a distinction between intra group debt and third party debt, nor does it make a distinction between cross-border and domestic debt.

In the preamble it is stated that MS have the option to apply an EBIT-test in an equivalent way, which however, is undefined. Net borrowing costs are defined in its broadest sense. MS have the possibility of granting the taxpayer the right to deduct exceeding borrowing costs up to a minimum of 3 million EUR, and to fully deduct exceeding borrowing costs for a stand-alone entity. Moreover, there is an optional application at group level if an entity is a member of group which may opt for tax consolidation.

Certain options are offered the MS in terms of possible escape clause options which can be applicable if the taxpayer is a member of a consolidation group for accounting purposes (IFRS or local GAAP definition).

–          Option (a): Fully deduct exceeding borrowing costs, if taxpayers can demonstrate that the ratio of equity over total assets equals or exceeds the group-ratio. A two percentage point deviation is accepted.

–          Option (b): Deduct some exceeding borrowing costs.

Options for carry forward/back are the following:

–          Infinite carry forward of restricted borrowing costs

–          Infinite carry forward of restricted borrowing costs AND carry back up to 3 years

–          Infinite carry forward of restricted borrowing costs AND carry forward of surplus EBITDA (interest capacity) for up to 5 years

The ATAD provides a grandfathering clause for certain loans, i.e. loans concluded before 17th June 2016 (remaining unchanged) and long term public infrastructure projects within the EU.

As an initial comment to the rule we believe that the ATAD largely follows the BEPS recommendation in Action 3. The EBITDA model is a simple model to “connect” taxable income to deductible financing costs.

Corporate groups should pay attention to the group ratio (balance sheet) and make sure to provide documentation for existing loans.

The EBITDA-provision does not one-to-one match any existing regime within the EU and accordingly it must be thoroughly assessed whether existing legislation in any EU MS can be considered a correct minimum implementation for the ATAD or if substantial changes are needed.

As described elsewhere the EBITDA- provision may face some constitutional issues in certain EU MS, since the German Bundesverfassungshof in its decision of 10th February 2016, has held the German Zinsschränke (very similar to the ATAD EBITDA-rule) to be in conflict with the net principle as derived from the German constitution. The case has been appealed, but if upheld, it would be difficult to apply such a rule in German law for domestic purposes, whereas a pure cross border application on the other hand would conflict with the TFEU.

Finally, it has been questioned whether the EBITDA-rule is consistent with primary EU law, as it may be argued that the rule does neither adhere to the counter-evidence rule nor the arm’s length principle (see e.g. Pieter van Os, EC Tax Review, 2016, p. 184-198). Clearly, this is only a problem insofar as a domestic implementation of the EBITDA rule targets cross border debt financing and not domestic debt financing.

Exit tax

The final ATAD introduces an exit tax in certain cross border situations on an amount equal to the market value of the transferred assets less their value for tax purposes. The exit tax is triggered at the time of exit in the following situations: Transfer of assets/business between a head office and a PE (or between PEs), or where a taxable entity transfers its tax residence.

The market value is the amount for which an asset can be exchanged or mutual obligations can be settled between willing unrelated buyers and sellers in a direct transaction. It is uncertain how the applied terms are in line with the arm’s length principle.

According to the provision the recipient MS, shall accept as an entry value the value used by the exit MS with respect to the exit taxation, unless this value does not reflect the market value. Unfortunately, the provision does not contain any safeguards against double taxation. In theory, the exit MS may impose an exit tax at a higher value than the market value since the directive is a minimum directive. Consequently, the recipient MS is not obliged to use this higher value as entry value.

The exit provision allows for a deferral via annual installments over five years. This means that longer deferral is not accepted. The deferral is allowed if there is an exit to another EU/EEA MS, but will be discontinued if:

  • The transferred assets are disposed of
  • The transferred assets are subsequently transferred to a third country
  • The taxpayer’s tax residence or its PE is subsequently transferred to a third country
  • The taxpayer goes bankrupt or is wound up
  • The taxpayer fails to honor installments over a reasonable period of time

Interest may be charged in accordance with the legislation in the MS. In addition the exit MS may require a guarantee if there is a demonstrable and actual risk of non-recovery.

The provision contains certain exemptions for temporary assets

Although very common among EU MS the exit tax is not part of BEPS but can be traced back to the CCCTB project. The provision is designed to reflect the comprehensive case law of the ECJ.

There does not seem to be room for stricter domestic legislation within the EU (ECJ case law), which in essence means that EU MS cannot introduce a stricter exit regime than the one prescribed in the ATAD provision.


One of the most significant parts of the ATAD is the introduction of a real GAAR (not like the PSD GAAR or the OECD PPT with limited scope). The GAAR is a back stop introduced to take over where the ATAD SAARs do not work. The GAAR targets all non-genuine arrangements domestically and in cross border situations and is intended to work as a supplement to the SAARs of the ATAD. In its design the ATAD GAAR closely resembles the PSD GAAR and it is designed to reflect the artificiality test of the ECJ. However, compared to the PSD GAAR the ATAD GAAR has a much broader scope.

Much can be said with respect to the scope of the GAAR, and surely, different EU MS have different experiences regarding the use of such provisions in their tax system.

The legal effect from its application is that arrangements etc. shall be ignored for the purposes of calculating the corporate tax, which shall be calculated by reference to substance in accordance with national law.

The legal effect of ignoring a transaction for tax purposes is not clear, in terms of what should be placed instead. Moreover, it is not clear how to interpret the reference to national law, i.e. what is applicable tax law in a given EU MS?

The legal requirements which should be fulfilled are the following:

–           “Arrangement or series thereof“ (an arrangement may comprise more than one step or parts)

–          Having been put in place for the main purpose or one of the main purposes of obtaining a tax advantage

–          That defeats the purpose or object of the otherwise applicable tax provision

–          Which is non “Non-genuine” (defined as not being put into place for valid commercial reasons, which reflect economic reality)

Even these requirements give raise to numerous questions. Just to mention a few:

  • How do you determine when an arrangement is put in place for the main purpose or one of the main purposes to obtain a tax advantage?
  • What is the purpose and objective of the tax rules?
  • Which reasons are valid (wording implies a value judgment)?
  • How do you determine economic reality?

These requirements are hard to apply in practice and very few guidelines are available. It is stated directly in the preamble to the ATAD that the GAAR is designed to reflect the ECJ’s ”artificiality test” when applied within the EU. It is not possible to assess whether this is correct, but it should be mentioned that it seems to be a bold statement since there continues to be uncertainty as regards the actual position of the ECJ with respect to different concrete situations. Future ECJ case law on the artificiality test should accordingly affect the interpretation of the GAAR. In our view the GAAR should be interpreted in light of the ECJ abuse doctrine and should not grow wild detached from this. In fact, despite differences in wording there seems to be little that separates the GAAR from the ECJ doctrine. If interpreted in accordance with the ECJ abuse doctrine it seems the practical scope is not that broad and that we can only rely on the limited guidance from Cadbury Schweppes, Thin Cap-cases etc. (physical presence, substantive activities, adhere to the arm’s length principle, acceptance of tax motives etc.).

Despite the potential broad scope of the ATAD GAAR tax authorities should carry the burden of proof. In practice the tax authorities should demonstrate that the above requirements are fulfilled, and then the taxpayers should try to prove contrary to this.

The ATAD GAAR is largely similar to the Principle Purpose Test of BEPS Action 6. In our view inspiration or maybe even interpretational guidance can be found in the relevant BEPS report.

In light of how far reaching consequences the GAAR might have and the wide scope it is unbelievable how little attention the GAAR has received. The GAAR will inevitably lead to uncertainty in general as well as with respect to the interaction with the SAARs. Moreover, there is a risk of different implementation, i.e. a risk which can result in as much as 28 different versions of the GAAR. Since there is no mechanism of mutual recognition or corresponding adjustment the GAAR may lead to double taxation. Hopefully, the scope of the EU Arbitration Convention will be widened or another tool for dispute resolution will be introduced.

Many business transactions must also be reviewed in light of the GAAR. In our view the actual impact can be boiled down to the question of how difficult it will be to pass the artificiality (genuine) test. Based on current case law of the ECJ in terms of this criterion, it might be only abusive transactions which are at risk. Clearly, it cannot be expected that all EU MS authorities will apply such an interpretation of the GAAR which is why from a practical perspective a wider scope is expected.

No direct guidance is available with respect to interpretation of the GAAR. It is seen that the final GAAR during the negotiation process was brought in alignment with the PSD and the OECD PPT. This must be applauded in general, although the parallelism is not fully certain. The ATAD GAAR is generally applicable – domestically and cross border and not only limited to tax treaties or intra group dividends.

Based on existing examples used to illustrate the PSD GAAR and the PPT, the following scenarios can be identified where some guidance is available:

  • Certain financial industry dividend stripping structures
  • Classic conduit/flow through structure – Are holding companies genuine? (Case C-6/16 Holcim)
  • Determining the location of production facilities
  • Determining the location of joint venture entities
  • Application of beneficial provisions through increase of ownership/shares. (E.g. increase from 9% to 10% or from 24% to 25%, new share classes etc.)
  • Mismatches not covered by other SAARs

CFC Legislation

CFC legislation formed part of the BEPS project as well as 2011 draft CCCTB directive. Accordingly, it is not surprising that the ATAD also include CFC rules.

The CFC rules in the final ATAD contain significant changes compared to draft version, and apparently it took a lot of effort to reach agreement on this matter. The result is a rather ungraceful compromise that includes two different models on how to include income from the CFC (i.e. a foreign entity or PE).

In order for the CFC rules to apply two basic conditions have to be fulfilled.

  1. The taxpayer should, by itself or together with its associated enterprises, control the entity. The control test is rather wide, as control is deemed to exist if the participation directly or indirectly is above 50 % of voting rights, capital or profits.
  2. Actual corporate tax paid should be lower than the difference between the corporate tax that would have been charged under the applicable corporate tax system of the taxpayer and the actual corporate tax paid. In effect, this amounts to a 50 % low tax threshold.

If these two conditions are fulfilled the taxpayer has to include the income of the CFC according to one of two different methods, depending on how the MS decide to implement the directive. The first method entails that the taxpayer shall include the following non-distributed income of the CFC (“tainted income”):

  1. Interest or any other income generated by financial assets
  2. Royalties or any other income generated from intellectual property

iii.            Dividends and income from disposal of shares

  1. Income from financial leasing
  2. Income from insurance, banking and other financial activities
  3. Income from invoicing companies that earn sales and services income from goods and services purchased from and sold to associated enterprises, and add no or little economic value

The IP category seems to have a broad scope while the category covering invoicing companies is uncertain in its scope.

With respect to the first method the income to be included should be calculated in accordance with the corporate tax rules of the state where the taxpayer is resident (losses of the CFC shall not be included in the tax base but may be carried forward and taken into account in subsequent tax periods). A MS may opt not to treat an entity or a PE as a CFC, if one third or less of the income is tainted income. Moreover, a MS may opt not to treat financial undertakings as CFCs, if one third or less of the “tainted income” comes from transactions with the taxpayer or its associated enterprises.

In addition, if the MS opts for the first method, the CFC rules shall not apply if the entity or PE carries on a substantive economic activity supported by staff, equipment, assets and premises, as evidenced by relevant facts and circumstances. However, concerning entities and PEs in third countries MS may decide to refrain from applying this exception.

The second method entails that the taxpayer shall include the non-distributed income of the CFC arising from non-genuine arrangements which have been put in place for the essential purpose of obtaining a tax advantage. In this regard an arrangement should or a series thereof shall be regarded as non-genuine to the extent that the CFC would not own the assets or would not have undertaken the risks which generate all, or part of, its income if it were not controlled by a company where the significant people functions, which are relevant to those assets and risks, are carried out and are instrumental in generating the controlled company’s income. As a result, the income to be included shall be limited to amounts generated through assets and risks which are linked to significant people functions carried out by the controlling company. Moreover, the attribution of income shall be calculated in accordance with the arm’s length principle.

According to both methods the income to be included shall be calculated in proportion to the taxpayer’s participation. Moreover, in both cases the CFC rule includes provisions that should mitigate double taxation in case of dividend payments from the CFC or a disposal of the participation in the CFC. Finally, it follows that the MS of the taxpayer shall allow a deduction of the tax paid by the CFC from the tax liability of the taxpayer (credit relief).

The CFC rule in the ATAD is expected to have significant effect on the corporate tax landscape in the EU, as 14 MS currently do not have CFC legislation in place. The compromise resulting in two different inclusion methods complicates the CFC rule and makes it hard to assess the implications in more detail. In this regard it is worth mentioning that MS can also decide to apply an entity-method and target the entire income of the CFC, cf. the preamble to the ATAD. Accordingly, at least three variations seem to exist concerning how to include income from the CFC. As a result the final CFC rules implemented by MS must be expected to vary substantially in scope and effect.

The BEPS report on action 3 recommends that jurisdictions allow an indirect ordinary credit relief for foreign taxes actually paid, i.e. taxes paid by the CFC itself as well as CFC tax assessed on intermediate companies (i.e. the situation where more than one jurisdiction applies its CFC rules to income of the same CFC). The CFC rules of the ATAD do not explicitly mention credit relief with respect to CFC tax assessed on intermediate companies, as the wording only addresses tax paid by the CFC itself (however, in the preamble to the ATAD it is generally stated that the rules of the directive should not only aim to counter tax avoidance but also avoid creating other obstacles to the market, such as double taxation). As more jurisdictions, inside as well as outside the EU, are expected to introduce CFC legislation in the coming years, the risk of simultaneous CFC taxation of the same income in two or multiple jurisdictions will most likely increase. Hence, there seems to be a serious risk of double taxation if the implemented CFC rules do not take into account the possibility of multiple CFC taxation when relief is granted.

Finally, with respect to the fundamental freedoms it appears reasonable to question whether the exception for CFC’s with substantive economic activity is enough to consider the CFC rules of the ATAD in line with the case law of the ECJ on abuse, which traditionally have only allowed a difference in treatment if the national measure specifically relates to wholly artificial arrangements. In addition, if MS opt not to apply the substantive economic activity exception to CFCs in third countries, it may be questioned whether the CFC rules are in line with the free movement of capital, as the control test may also be fulfilled only through economic control (entitlement to more than 50 % of profits).

Mismatch arrangements

The final ATAD includes a new approach with the aim of combatting hybrid mismatches, which is more in alignment with the OECD BEPS recommendations than the proposed source state superiority initially proposed.

The provision now aims at the outcome of any mismatch which may result in ”double deduction”, or “deduction without inclusion”.

Hybrid mismatches are defined as situations between a taxpayer in one MS and an associated enterprise in another MS OR a structured arrangement between parties in MS where double deduction or deduction non-inclusion is caused by differences in the legal characterization of a financial instrument or entity. The latter term of structured arrangements is not defined. The definition of associated enterprises differs depending on the context. When it comes to hybrid instruments the following applies: 25% ownership of capital, voting rights or profit participation (directly or indirectly). When it comes to hybrid entities the following applies: 50% ownership of capital, voting rights or profit participation (directly or indirectly)

–          Double deductions are defined as a deduction of the same payment, expenses or losses occurs in both MS. The legal effect in cases of double deductions is that the deduction shall be given only in the MS where such payment has its source, the expenses are incurred or the losses are suffered and in another MS.

–          Deduction non-inclusion is defined as situations where there is a deduction of a payment in the MS in which the payment has its source without a corresponding inclusion for tax purposes of the same payment in the other MS. The legal effect is that the MS of the payer shall deny the deduction of such payment.

As an overall comment it is not surprising that the ATAD contains an anti-hybrid/mismatch provision. Such provisions have become increasingly popular in recent years. Despite this, only a minority of EU MS have introduced such provisions. The ATAD as a consequence will significantly change the tax landscape in such MS.

In terms of the actual design of the provision it should be noted that it only applies to intra-EU mismatches. It is directly stated in the preamble that further work on third country situations has to be carried out

With respect to the scope of the provision it should apply to many debt-equity hybrids, REPOs, group contributions, silent partnerships, etc., but not e.g. tax-credit arbitrage.  The current provision is not applicable to lower or higher tier mismatches through intermediate companies, although MS may introduce such a broadened scope (which is already seen in a number of EU MS domestic legislation). The ATAD provision is furthermore not applicable to ACE (NID) regimes since (could be referred to as institutional hybrids) there is no payment taking place.

Generally speaking linking rules may constitute a violation of fundamental EU freedoms. This question has been analyzed by Jakob Bundgaard: “Hybrid Financial Instruments and Primary EU Law Part I and II” in European Taxation 2013, p. 539 et seq. and p. 587 et seq. The author concludes that domestic linking rules within the EU seem to conflict with EU law. Despite this potential conflict no analysis seems to have been made regarding the use of cross border linking rules in the ATAD. There is one significant difference in the fact that the ATAD is secondary EU law. Although secondary EU law may in principle conflict with primary EU law, it seems that the likelihood of being overruled by the ECJ is lower than purely domestic rules.

Preliminary comments

It is still too early to make an impact assessment of the final ATAD. However, some general remarks can be made.

Overall, it seems expedient that the Commission has taken on the task of implementing the BEPS recommendations in a coordinated way within the EU. However, as the ATAD is a minimum directive, it gives MS some leeway to design and apply their anti-avoidance rules, as they would like. As a result, the scope of application and the effect of the anti-avoidance rules across MS may end up differing considerably, and the objective of securing a coherent and coordinated transposition of the BEPS measures into MS national tax systems may therefore be hard to reach.

Another issue is whether the anti-avoidance rules of the ATAD are fully in line with the requirements following primary EU law, in particular the ECJ’s case law on the fundamental freedoms with respect to abuse. In particular, the EBIT-rule, the GAAR and the CFC rule seem to raise such concerns.

One question that comes to mind is whether the ATA-Directive fits its policy objective. Obviously, the new provisions (not taking into account the variations in domestic implementation) will strengthen anti- tax avoidance measures in most EU MS. Consequently, many aggressive tax planning structures are at risk. However, on the other side, the ATAD seems to target more than aggressive tax planning, as a number of typical business transactions/structures are at risk of being affected by the ATAD (over BEPS-ification). This means that transaction costs may occur in the process of demonstrating why the rules should not apply to such transactions/structures. As a worst case, double taxation and new disparities may occur due to the nature as a minimum directive.

From a practical perspective, several structures should be reviewed in light of the ATAD.


Leave a Reply

Your email address will not be published. Required fields are marked *