Some European views on the Swedish tax system

a. Contents of this article

In the below contribution I have summarized some of the key findings of those insights.

Firstly, I will share some of my thoughts on the Swedish coupon tax in connection with the Sofina case and some insights from a Dutch point of view. Secondly, I will discuss how the Swedish implementation of the 30% EBITDA rules interact with the group contribution regime. This I will then test against the framework set by the Groupe Stéria case and share the Dutch perspective on this matter. Thirdly, I will elaborate on the taxation of funds on an investment savings account against the background of the combined cases of X, Miljoen and Société Générale, as well as from a Dutch standpoint.

I will conclude with some thoughts on what is yet to come and what would be the EU law aspects thereof.

In all the below I will give a brief description of the relevant EU provisions or case law. However, when it comes to the relevant Swedish law, I will only stick to the minimum that is relevant for clarifying the point I want to make, as I am not yet in the position to write anything meaningful on the Swedish tax law. Or, as the Dutch would say: “Who is the snack bar here; you or me?”3

b. European testing framework

Where I will test the Swedish law against the rules laid down in the European primary and secondary law, I will apply the standard framework4 as established by the European Court of Justice when testing the compatibility of a certain situation with the freedoms as laid down in the TFEU, and the prohibition of discrimination. With the risk of stating the obvious, I will briefly go into this framework so that the below should be accessible for all readers.

• Firstly, it should be determined if a taxpayer has access to the TFEU. This condition is met if the presented case concerns an economic activity and a cross-border situation.5

• Secondly, it should be tested whether a measure is discriminating:6

  1. two comparable situations

  2. treated differently

  3. whereby the cross-border situation is treated disadvantageous

  4. and this is not the result of a disparity

  Or constituting a breach with one of the freedoms laid down in the treaty:7

  1. Free movement of goods

  2. Freedom of persons

  3. Freedom of establishment

  4. Free movement of services

  5. Free movement of capital

• Thirdly, it should be assessed if such restriction can be justified. Accepted justifying grounds are:

  1. Effectiveness of tax audits8

  2. Coherence of tax system9

  3. Fiscal territoriality10

  4. Combating tax avoidance11

• Lastly, if a restriction can be justified, the measure at hand should be suitable and proportional, meaning that the breach is required and should not go further than required to achieve the ground that justifies the measure.12

I understand that my findings may appear to criticize the Swedish tax system. However, I ask the reader to keep in mind that it is generally of more academic value to identify which parts of a tax system are incompatible with EU law, than to identify what is in line with the requirements of our European (tax) rules.

a. Sofina case

A quite recent but rather interesting case is the Sofina case, ruled by the European Court of Justice on November 22, 2o18.14

This case concerned the French dividend withholding tax, and particularly in situations where the recipients of those dividends were in loss making positions. In the Sofina case, three Belgian loss-making entities received French dividends. On these dividends, 15% French withholding tax was levied, in accordance with the Belgian-French double tax treaty. As the companies were loss-making there was no Belgian income tax against which they could credit the French withholding tax, which made the withholding tax an irrecoverable expense.

However, if the Belgian companies would have been French tax residents, they would have had the possibility to postpone taxation until the moment they get profitable again. If this would never happen, the tax would not fall due at all.

The court ruled that this was a prohibited restriction on the free movement of capital for which no justification ground existed.

b. Coupon tax regime

The Swedish coupon tax regime effectively works similar as the French withholding tax regime.

In accordance with paragraph 4 of the Coupon tax act Swedish tax residents are not subject to coupon tax. Instead, they are taxed for the dividends they receive through their income tax.

If we would compare the relevant considerations in the Sofina case with the Swedish coupon tax regime, this would lead to the following outcome:

As per the above the Sofina argumentation can be applied without material differences to the coupon tax regime.

c. The Dutch perspective

In the Netherlands, withholding taxes are both for national and for foreign recipients, a final tax. Although they can be credited about local (personal or corporate) income tax, they can never result in a zero balance.

As a result, the Sofina case should not have any direct impact in the Netherlands.

d. Conclusion

As the argumentation of the European court on the French withholding tax regime in the Sofina case can be applied similarly to the Swedish coupon tax regime, its conclusions could potentially apply similarly as well.

On that basis it should be concluded that the levy of coupon tax in case of a loss-making recipient is a forbidden restriction of the free movement of capital for which no justification ground exists and that such loss-making entities receiving Swedish dividends should be entitled for a refund or exemption until they are generating profits again.

Quite recently, the Swedish department of finance has recognized this when they published a proposal to amend the coupon tax regime as from 2020.15 It is questionable though, if they did the right thing by publishing this proposal now, with an effective date of January 2, 2020. Many taxpayers who are in a Sofina situation (i.e. loss-making in their home state, but receiving Swedish dividends on which coupon tax was withheld), may now feel strengthened in their position and claim complete refunds.

They would be entitled to do so, from a European perspective, because the coupon tax was levied based on legislation that caused a breach with European law. In the past, the Dutch finance department has worked with the instrument of “emergency measures” to combat those unwanted interim periods of uncertainty.16 These emergency measures imply that the minister of finance publishes a decree in which the legislation is changed directly to bring it in line with the European requirements, but where he keeps the right to fill in the details of the new rules at a later stage.

A last comment, not necessarily related to the Sofina case, but in fact related to the revision of the coupon tax that is now happening in Sweden. Although arguably of little relevance for the everyday tax practice, charitable foundations that would have been exempt from Swedish income tax if they were a Swedish resident, should also be exempt from coupon tax if receiving dividends with a Swedish source. There is no reason that I can think of, why Swedish and non-Swedish charitable foundation should be treated differently.

a. Groupe Stéria

The Groupe Stéria case18 is a case of which the importance can hardly be overestimated. It basically challenges some of the key features of virtually all tax grouping regimes in Europe.

The appellant was heading a French tax group but also held subsidiaries in other member states which it could not include in the tax group. According to French law, dividends from these other subsidiaries were 95% exempt where the non-exempt 5% were to represent a fixed amount to cover ant costs and expenses that a parent company would bear for its subsidiary. If the subsidiaries would have been part of the same tax group as their parent company, the dividends would have been fully exempt. However, as the subsidiaries were not French they could not elect to be part of this tax group and therefore their French shareholder could not benefit from a 100% dividend exemption.

The court ruled that this was a prohibited restriction on the freedom of establishment for which no justification ground existed.

b. Swedish implementation of the 30% EBITDA rule

In short, the 30% EBITDA rule as implemented in Sweden is not very surprisingly limiting (net) interest deduction to 30% of the EBITDA. The part that is interesting from an EU perspective, is that if two entities are eligible for group contributions, they are also eligible to exchange interest income and expenditures for the purposes of the 30% EBITDA calculations.19 However, in order to be eligible for group contributions, you need to be a Swedish tax resident.20 So if a Swedish company with a net interest cost has a group company with excess interest income, it could be caught by the 30% EBITDA rule if the group company was foreign. However, if the company was a Swedish resident and would be eligible for group contribution, the entity could choose to transfer some of the excess interest income so that it’s group company would not be caught by the 30% EBITDA rule.

If we would compare the relevant considerations in the Sofina case with the Swedish coupon tax regime, this would lead to the following outcome:

As per the above the Groupe Stéria argumentation can be applied without material differences to the coupon tax regime.

c. The Dutch perspective

I believe that the Dutch legislator would have been lucky if his problems in relation to the fiscal unity, were limited to the 30% rule. One of the emergency measures mentioned in 2.d. actually related to the fiscal unity and was a direct consequence of the Groupe Stéria case. Prior to the Groupe Stéria case, and the emergency measures that were to follow, a fiscal unity for corporate income tax purposes effectively meant that the entities united in this fiscal unity were treated as if they were one. This goes way further than the mere exchange of tax losses, but also means that no transfer pricing documentation would be required (as no intercompany transactions were recognized), that interest deduction limitations applicable to intercompany interest do not apply as no intercompany interest was recognized, etc.

It is quite remarkable that the emergency measures that were to deal with the consequences of Groupe Stéria did not say a word about how the 30% EBITDA rule would be effectuated in that light – while the scope of those measures was known at the date of publication. To date, also in Dutch literature and legislative working papers, no clear line is drawn on this matter.

It is expected that the Dutch fiscal unity regime will soon be restructured more thoroughly, potentially even allowing foreign entities to be part of such fiscal unity and have them treated as if they were permanent establishments. Then, also the issues related to the 30% EBITDA should be solved. However, until that day, we can expect quite some interesting discussions on this topic in the Netherlands.

d. Conclusion

As the argumentation of the European court on the French tax grouping regime in the Groupe Stéria case can be applied similarly to the Swedish implementation of the 30% EBITDA rule, its conclusions could potentially apply similarly as well.

On that basis it should be concluded that the disallowance of transferring excess interest income from abroad to Swedish group companies with a net interest cost is a forbidden restriction of the freedom of establishment for which no justification ground exists and Sweden should accept the transfer of foreign excess interest income in the formula for calculating the interest deduction limitation in Sweden.

a. The combined cases of X, Miljoen and Société Générale

In 2015, the European court of Justice ruled on a case dealing with two Belgian individuals (X and Miljoen) and one French resident company (Société Générale) that all filed an objection against the dividend withholding tax levied in the Netherlands.22 Their complaint was in short that although the 15% dividend withholding tax was the same for Dutch tax residents as well as for foreign tax residents, foreign residents could be in a worse position as they could have the 15% as a final levy, where Dutch tax residents could credit it against income tax due. So in a way this is pretty similar as the Sofina case. However, as opposed to the Sofina case, here withholding tax was levied both in the national and in the cross-border situation. The complaints related to what happened after that. Because if X and Miljoen would have been Dutch tax residents, they would have been taxed on a flat rate basis – just as with the ISK – and be able to claim the actual withholding tax as a prepayment of income tax. Then, depending on the average returns, value of the investments and any debts of that person, the effective tax may be lower than tax collected by withholding.

The court ruled that this was a prohibited restriction on the free movement of capital for which no justification ground existed.

b. The taxation of Investment Savings Accounts

The ISK regime effectively works similar as the “box 3” taxation that Miljoen and X referred to in their case: a flat rate based taxation calculated on the value of ones assets and multiplied by the standard tax rate. The difference between the ISK taxation and the Dutch box 3 taxation lies in the fact that the ISK is an actual banking product so income needs to be on a specially designated savings account in order to benefit from the flat rate taxation. In the Netherlands, all ones assets qualify for this taxation.

If we would compare the relevant considerations in the cases mentioned above (taking Miljoen as an example) case with the ISK tax regime, this would lead to the following outcome:

As per the above the Miljoen argumentation can be applied without material differences to the ISK regime.

c. The Dutch perspective

As the Miljoen case is Dutch, the Dutch angle to this is more easy than in the Groupe Stéria and the Sofina cases. However, the impact is much bigger than compared to the situation with the Swedish ISK. This is because of the fact that the Dutch “box 3” treatment follows automatically from the law. So in case one has savings or listed shares, they are automatically taxed based on the rules of Box 3, without having a specifically designated banking product for that purpose.

d. Conclusion

As the argumentation of the European court on the Dutch box 3 regime in the Miljoen case can be applied similarly to the ISK taxation, its conclusions could potentially apply similarly as well.

On that basis it should be concluded that the coupon tax withheld on dividends should be refunded to foreign shareholders, to the extent that the recipient of those dividends:

• could have used an ISK for those shares if he would have been a Swedish tax resident, and

• the effective taxation based on the ISK regime was less than the Swedish coupon tax withheld by the foreign recipient of those dividends.