Exploring the varying approaches taken by EU Member States to the implementation of DAC6

This article was originally published on STEP . Read the article here.  

One would be forgiven for having overlooked the second birthday of EU Council Directive 2018/822 (DAC6), particularly as, in the majority of Member States, there are no longer any DAC6 filing and reporting deadlines in 2020.

As a result, it may be tempting to take one’s eye off the ball. However, although filing and reporting may have been delayed, potentially reportable arrangements continue to accumulate; DAC6 remains applicable to relevant arrangements from 25 June 2018. As such, those who have been analysing new arrangements from 1 July 2020, rather than amassing a second backlog of cases, could find themselves at an advantage. Austria, Finland and Germany have not postponed their deadlines.

It may also be tempting to over-report, whether out of an abundance of caution or because it is easier. However, if nothing else, it may do a disservice to clients to suggest that they are consistently involved in potentially aggressive tax-planning arrangements. To report when one is not strictly supposed to could also contravene data protection and professional secrecy obligations.

The better alternative would be to dust off the decision tree in the first article in this series [1] and apply it methodically in light of the local law, explanatory notes and guidance that are now emerging.

Festina lente

Let us suppose that three friends are each looking to invest a portion of their wealth. They reside in Belgium, the Netherlands and the UK and, with the involvement of local advisors, they each purchase an investment-linked insurance policy on 1 October 2020 from a Luxembourg insurer. Given the breadth of scope of DAC6, do the policies need to be reported under DAC6? The reporting of established planning such as this should be the exception rather than the rule, but is the analysis the same across Member States?

Our starting point is whether an arrangement exists. For our UK friend, an arrangement ‘includes any scheme, transaction, or series of transactions’. [2] According to the Netherlands guidance, [3] it could include ‘a transaction, action, agreement, loan, arrangement, or a combination thereof’. The Belgian guidance tells us that a life policy in particular could be an arrangement. [4] Luxembourg follows the wording of DAC6.

Is the arrangement cross-border? Though the answer appears obvious (if participants are in separate Member States, the arrangement is cross-border) local rules differ. The Luxembourg insurer might well consider itself a participant as, without its involvement, the policies could not exist. The UK guidelines [5] ask us to step back and contemplate whether an arrangement concerns more than one jurisdiction at all, even if there could be participants in each. For example, investment in a foreign collective investment scheme will not always concern multiple jurisdictions.

Interestingly, the Belgian view is that a foreign insurer will not be a participant if it does no more than issue a policy, [6] in which case the Belgian analysis is complete and there is no Belgian report. In the Netherlands, whether a person is a participant ‘will depend on the circumstances of the case’. [7]

If there is indeed a cross-border arrangement, the next question is whether a hallmark is present. In the previous article, we looked at two:
 

• arrangements having substantially standardised documentation; and
• arrangements having the effect of converting income.

Neither seems directed at the policies, but all hallmarks must be considered. The UK [8] and the Netherlands [9] are aligned: documents requiring customisation to a client’s circumstances are usually out of scope. In the same vein, Belgium excludes situations requiring personalised advice, [10] a stance mirrored in the explanatory notes to the Luxembourg draft implementing law and the Luxembourg insurance industry FAQs. [11] The latter take the view that standardised policy application forms and general conditions are also not the focus of DAC6. It would be perverse that the simpler and more legally prescribed a set of standard conditions, the more likely that the arrangement is aggressive.

Conversion of income is more nebulous. What constitutes conversion and whose income must be converted? As suggested in the Luxembourg FAQs, there is unlikely to be an interest for a taxpayer in the conversion of a third party’s (the insurer’s) income on policy-linked assets. The position could differ if there was a pre-existing income stream, but with insurance something more would be required, as ownership is lost on the transfer of the premium to the insurer and any future entitlement is only to policy value.

Unless an alternative hallmark applies, or there is a broader reportable arrangement, our three clients (or their advisors) are, so far, unconcerned by DAC6. However, if there is a hallmark, one must consider whether DAC6’s main benefit test (MBT) applies. This asks whether the, or a, main benefit to be expected from the arrangement is a tax advantage. Remember that not all hallmarks require the MBT and some hallmarks (notably category D) require little or no tax element at all.

The first question is: which tax? The tax advantage should relate to one or more of the taxes in Council Directive 2011/16/EU of 15 February 2011, [12] although, according to Luxembourg [13] and the UK, [14] among others, if the advantage is derived outside the EU, it is still relevant for the MBT. Member States have also ventured definitions of tax advantage. These variably include reliefs, repayments, avoidance, reductions and deferrals. Whether the tax advantage is the main benefit will depend on an assessment of the facts and circumstances.

Helpfully, guidance in some countries acknowledges that tax authorities are not interested in recapturing information on tax advantages that are in accordance with legislative intent. This is the view in Luxembourg, for example. [15] The Netherlands guidance defers to the UK, where a tax advantage in line with legislative intent is not a tax advantage at all. [16] In Belgium, we are only told ‘the mere application of a preferential (foreign) tax regime does not automatically satisfy the main benefit test’. [17]

For example, our UK friend will benefit from tax deferral and subsequent withdrawals from the policy potentially with no immediate tax charge. However, the chargeable events regime for insurance contracts provides expressly for this treatment, in which case the tax advantage in question is not relevant for the MBT. In Belgium, policies are also taxed in an established way. If they are not taxed favourably, there is no advantage to assess for the purpose of the MBT. In the Netherlands, the policy is within Box III and is taxed on an ongoing basis. There, the main benefit of the arrangement will necessarily be another of the advantages for which insurance of this type is known.

Parting thought

DAC6, in accordance with the subsidiarity principle, sets out to achieve an objective that ‘cannot sufficiently be achieved by Member States’. That objective is a good one. However, between the drafting of DAC6 and its implementation in national law, there is a risk that something will be lost in translation. As such, it will pay dividends to know how neighbouring countries are approaching the same questions.

Ultimately, intermediaries of all varieties are likely to encounter at least some reportable arrangements. But they should not forget that, far from constituting an additional anti-avoidance regime or an obligation to seek out the least (tax) efficient in any series of investment options, DAC6 is directed at potentially aggressive schemes that threaten Member State tax revenue. As such, it should be a shield for genuine and established planning and reserve the sword for the remainder. [18]

Simon Gorbutt Director – Regional Head, Wealth Structuring Solutions Lombard International Assurance S.A