Exit taxes on corporate relocation
Andrew Parkes, National Technical Director, discusses dual holding structures and departure taxes, following recent issues impacting Unilever both in the UK and Netherlands.
Andrew’s article was published in Taxation, 8 December 2020, and can be found here.
To those of us who did our tax technical training with HMRC – or the Inland Revenue as was – the case of Van den Berghs Ltd v Clark  19 TC 390 was required reading and, if the right question came up in the exam, four marks could be achieved, helping towards that magical 70% pass.
Now, one of the corporate ‘descendants’ of Van den Berghs, Unilever, may have to fight another case through the courts if the Dutch parliamentary opposition gets its way. The problem (for want of a better word) is the dual holding structure that Unilever has in the UK and the Netherlands. This was causing commercial issues for the group, for example making acquisitions and demergers more complicated, and it was decided to unify the structure with a single holding company based in the Netherlands. But there was a slight snag: the company’s shareholders disagreed with this decision and it was scrapped with a proposal to move to a single holding company in the UK being announced in June 2020. However, this attracted the ire of the Dutch opposition Green-Left party which proposed draft legislation in an attempt to stop the move.
The original legislation applied to any group with a turnover of more than €750m (this now seems to be the threshold of choice for new rules to apply to multinational groups), which sought to relocate its headquarters to a country that does not impose withholding tax on dividends, or does impose withholding tax but treats any pre-existing profits on entry as capital (and therefore not subject to withholding). Groups that fall within these parameters would face an exit tax of 15% based on undistributed profits and reserves.
It could not have been any clearer that this was aimed at Unilever if it had been called the ‘keep Unilever from leaving’ bill. However, it has failed in its aim because although Unilever estimated that, if effective, the legislation would cost £9.9bn (€11bn) it ploughed on with its move. The company obtained agreement from the UK’s High Court for the merger on 2 November and this took place on 29 November.
The €750m limit also caused problems as regards state aid because it only applied to a small subset of companies and was therefore selective in its application. In what appears to be a cosmetic attempt to make the law more likely to pass muster, the proposal has been widened to include any company leaving the Netherlands. This also means that it has a new starting date of 18 September 2020. However, there is an exemption for the first €50m of reserves, which will take most companies out of scope, but may bring the proposal back within state aid laws because it is still government support on a selective basis. Indeed, since the original proposals were introduced in the summer, the Green-Left party has made a number of what can only be described as vain amendments to its proposal, including ‘expanding’ the scope to make it legal, some of which I will discuss in this article.
Exit taxes have been through the Court of Justice of the European Union (CJEU) on a number of occasions, with the Dutch at the heart of two of the major cases, N v Inspecteur van de Belastingdienst Oost/kantoor Almelo (C-470/04) and National Grid Indus BV (C-371/10). These cases have allowed an EU-compliant exit tax to be designed and indeed required by the Anti-tax Avoidance Directive (2016/1164).
The CJEU cases were considering whether the exit taxes infringed the EU’s fundamental freedoms. Member states’ domestic legislation must not infringe these freedoms or, if they do, any infringement must be justified by an overriding need of public policy and, perhaps most importantly, not be disproportionate. It is at this hurdle of proving that infringements are not disproportionate that member states most often fall.
The two freedoms we need to consider in the present circumstances are the freedom of establishment (TFEU, Art 49) and the freedom of movement of capital (TFEU, Art 63).
The freedom of establishment allows any EU person to set up in business, including by way of a branch or a subsidiary, in any other member state. This stops a member state from introducing legislation that either discourages people from other EU member states from setting up a business and the like in the first state or from discouraging their own people from setting up businesses in another member state.
Infringements and justification
The proposed Dutch exit tax does not apply to Dutch companies moving their place of effective management within the Netherlands or merging with other Dutch businesses, but it does apply to cross-border moves and mergers. As such, it is clearly aimed at dissuading a Dutch company from moving its business out of the Netherlands and therefore infringes the freedom of establishment.
Infringements are justified for overriding reasons in the public interest, but they must be appropriate and not go beyond what is required to obtain that objective. It is at this last hurdle that national laws often fail because, although they may be justified and appropriate, they go beyond what is required. The Green-Left party has clearly taken note of this because the lack of deferred payments was what caused the Dutch corporate exit tax to fail in the National Grid case – and led to the UK’s own exit charge being amended. One of the amendments the party has made to its original proposal is that the tax is automatically deferred until a dividend is paid.
The CJEU has found that exit taxes are justified because of the ‘balanced allocation of powers of taxation’ between member states based on the ‘principle of territoriality linked with a temporal component’. In layman’s terms, because a company will benefit from the infrastructure of a country and the like while it is resident there, that country is allowed to tax any gains that arise during that time. It does not have to allow some other country to reap the taxation rewards of those gains following a move of residence.
Generally, as might be expected, exit taxes apply to unrealised gains (such as shares held in subsidiary companies). However, this Green-Left exit tax applies to profits that have already been taxed. Unilever has made its profits and they have been subject to Dutch corporate income tax. The net amount is then sitting in the accounts and is available for distribution. The Dutch government has already had its bite of the cherry.
This, then, is a fundamental difference between current exit taxes and the latest Dutch proposal. The rationale for the balanced allocation of taxing rights for the capital gains exit taxes can be seen. If an asset grows in value because of advantages provided by a country, it is, as mentioned, reasonable for it to insist that it receives a cut of the increase in value. However, for income profits, the state has already had its share by assessing the profits to income tax. The balance has already been established and honoured; here it is being unfairly tipped towards the Dutch, whether the tax is paid now or deferred.
An argument could be made that a move within the Netherlands and one to the UK are similar. This is because if Unilever were to move within the Netherlands and subsequently paid a dividend upon profits earned at its previous Dutch location, that dividend would be liable to Dutch withholding tax. Under the proposed law, the dividend paid from those earlier Dutch profits would also be liable to Dutch withholding tax if the dividend is paid after a move to the UK. However, the situations are not the same. A move to the UK causes the exit tax to become due immediately. Although the tax is deferred until a dividend is paid, the liability will sit on Unilever’s balance sheet which may affect its value and ability to borrow. Whereas a move within the Netherlands leads to no such problems.
It is understood that the original proposal had a provision to allow a refund or credit of the withholding tax to the recipient if payment of the exit tax is deferred until an actual dividend is paid. However, there was no such credit or refund ability if the tax is paid ‘up front’. This would seem to have made the tax disproportionate because it denied a credit or refund in some circumstances with no apparent justification. This has now been removed.
A step too far
A further consideration is that the proposed law relates to dividends. Both the freedom of establishment and the freedom of movement of capital can apply to the payment of a dividend. Which one depends on the purpose of the law. Is it designed to apply to companies that exert influence over the payer? If so, it is the freedom of establishment and, if not, then the freedom of movement of capital.
Here, we have already engaged the freedom of establishment because the law will apply to any company trying to leave the Netherlands. Therefore, if the CJEU applies this to the payment of the dividend, even though the law is not specifically aimed at shareholders who have influence over the company, (as the freedom of establishment is already engaged) then it is likely to bring the decision in Eqiom (C-6/16) into play.
This was a case involving France where a French subsidiary paid a dividend to a Luxembourg holding company, which was owned by a Cypriot company that was in turn owned by a Swiss parent company. The French law denied the benefits of the Parent Subsidiary Directive on the basis that the ultimate beneficiary was outside the EU.
The French law was aimed clearly at combating fraud and tax evasion and must, therefore, only be aimed at stopping wholly artificial arrangements that do not reflect economic reality, the purpose of which is to obtain a tax advantage. Therefore, a general presumption of abuse cannot justify the infringement of a fundamental freedom.
The Dutch proposal is now aimed at all companies moving to a country that does not levy a dividend withholding tax or treats any pre-existing profits on migration as capital (thus not subject to the new country’s withholding tax) when a company enters the country. It is, therefore, clearly aimed at stopping what the EU calls ‘tax evasion’, but the rest of us may call avoidance – the movement of a company to a country so that it does not have to pay withholding tax upon its dividends. As an aside, if the difference between avoidance and evasion is already blurred in the minds of many politicians and the public, having the EU call both evasion does not help at all.
However, many companies move for commercial reasons as well as tax ones. The Unilever move is to simplify its structure and to assist it with takeovers and/or demergers. This is clearly not artificial or solely tax driven as a measure that restricts the freedom of establishment requires, so I would expect the CJEU to strike it down.
As mentioned, the freedom of movement of capital may also apply where a shareholder does not exert influence over the company paying the dividend.
That way doesn’t work either
If the CJEU does consider the freedom of movement of capital, then it will be infringed. This is because the Dutch measure will clearly dissuade investors from outside the Netherlands from investing in companies within that state since they may face extra taxation (due to this exit tax) than they would if they invested in a company in another country.
The affected investors can be both inside and outside the EU and, therefore, will apply to UK investors after Brexit or if the Dutch proposal is amended to cover just non-EU investors (in an attempt to not infringe the freedom of establishment).
However, the tinkering by the Green-Left party will be of little effect because the justifications and proportionality issues apply to the freedom of movement of capital as they apply to the freedom of establishment. This means that the comments of the CJEU in the Eqiom case also apply here. Because the Dutch proposal is aimed clearly at what the EU calls tax evasion, it should only apply to wholly artificial transactions that do not reflect economic reality and are for a tax advantage.
Again, Unilever’s move is not wholly artificial. Indeed, it is not artificial at all, it will reflect economic reality and the withholding tax saving is not the only reason (I don’t think anyone would try to argue that it was not a factor). Because the Dutch proposal is for a blanket provision, it is not proportionate and should be struck down.