Radical change on the way for Dutch royalty conduits Newly introduced Dutch ‘substance’ requirements and an agreed change to the EU Royalty and Interest Directive may bring radical change to the way Dutch royalty conduits will operate in future, writes Jos Peters, the Senior Tax Partner at Merlyn International Tax Solutions Group Introduction The usual practice to benefit from the Dutch tax treaty network in the area of reducing or even avoiding foreign withholding taxes on royalty payments, has always been to set up Dutch BV as a royalty conduit, whereby the Dutch entity was managed by a Dutch fiduciary services provider (often called a ‘trust company’). The trust company charges a fixed fee for hosting the entity and also a fixed fee for the statutory director(s) it will make available. On top the trust company charges on a time spent basis what it does for the entity such as bookkeeping, preparing annual accounts, filing the annual corporate tax return but also providing legal and tax assistance. In very many cases the latter two categories have come from outside advisers (experts). It is a ‘public secret’ that the Rolling Stones, U2 and Tina Turner use such Dutch royalty conduits to receive almost tax free income from their performances worldwide. Also football players, tennis players, race car drivers, motor cycle racers and ice hockey players are known to use this traditional ‘Dutch route’ for their image rights. But the term ‘royalty’ in tax treaties covers a lot more: all legal types (for the licensing of a patent, a trade mark, a trade name), authors’ rights, but also payments for the rental of equipment, (airplane leases, bare boat charters…), the rendering of technical services and consultancy. Changes have already been announced, there are more to come This old royalty conduit model is almost outdated, however. Strangely enough, it is still often used because providers hate to tell clients that their solutions are losing tax and legal efficiency, or they may not even know that this is the case because legal and tax matters are usually boarded out by trust companies to external parties such as the Big Four for tax advice or the international law firms, either for tax advice or for assistance with contractual matters or for both. In my view there is (risk of) a major gap here: the trust company may not know that the changes in tax rules will affect the structures it operates for clients and the outside lawyers and tax advisers usually wait with advising and assisting till the phone rings and questions are asked. But the phone does not ring because the trust company is not aware of the (upcoming) problem. This leaves ample room for matters to fall between the ship and the shore, which – increasingly – implies the risk that a royalty conduit structure that was OK three years ago, is now no longer OK, but nobody signals this. Several years ago I have already started to offer my clients a different and much better solution. On our website www.merlyn.eu in the ‘publications’ section an article can be found that I wrote in late 2010 about this alternative structuring of a royalty conduit entity. I suggest the reader, for a full understanding of how the changes to the royalty conduit sector are playing out now, three years later, to read this article (written for the Euromoney Corporate Tax Handbook 2011) before continuing. My alternative back then, to use a Dutch royalty conduit owned by a third party, was already advantageous because it eliminated the effects of the so-called transfer pricing rules, always an expensive part of setting up a Dutch royalty conduit entity owned by the client himself. Should the entity be owned by the Dutch service provider, instead of only being hosted and managed by him, there is no need for a benchmark study and an advance tax ruling, so this set-up will eliminate an upfront expense of some €15K. It also eliminates the sometimes very costly outside legal and tax advice: the client no longer needs this (the legal and tax issues of the company in use are no longer his but the provider’s worry since the provider now also owns the entity) and the provider must have this knowledge in-house if he has sufficient working experience with this alternative business model. Background on why changes are seeing the light of day at speeds never seen before Recently, under pressure from the many press publications about the Dutch royalty structures employed by Yahoo, Starbucks, Apple, Google and the like, in their ‘Double Irish – Bermuda’ set-ups (which all heavily rely on a Dutch intermediate entity between the two Irish ones) many changes have been made or announced, by the Netherlands itself, by the European Commission and by the OECD, with the aim to counter the ‘immoral tax avoidance’ that takes place via the Netherlands, by these big multinationals. The Netherlands To start with, the Netherlands introduced new ‘substance’ rules in its Corporation Tax Act in September last, effective 1/1/2014. The country already had the strictest substance rules for its own legal entities, introduced in 2001. The main new challenges for 2014 and onwards are: 1) Advance rulings are no longer available for standalone royalty conduits; this risks of course a dispute with either the Dutch tax administration on what gross margin a Dutch royalty conduit entity should make in any given case, or a similar dispute with the foreign tax inspector where the royalties originate if the Dutch conduit entity is ‘related’ to the entity that pays the royalties; 2) If the Dutch tax authorities should find out, from several new questions added to the Dutch corporate tax return, or upon a tax audit that a structure falls short of the Dutch substance rules (and in case one does not have an advance ruling this could happen any time) it will send letters out to the foreign tax inspectors who deal with the royalty paying entities, informing them thereof. This will likely bring the roof down in the royalty paying countries and the tax authorities there will see the letter from the Dutch tax authorities as a sign that they can still in retrospect charge the royalty taxes they forewent in the past, during at least a five year period, with fines and interest... 3) Royalty conduits need to maintain a substantial amount of capital; this point was unclear until now (there was only a strict rule for interest conduit entities) but the ruling team has published a ‘safe haven’ approach for the royalty sector: the Dutch entity, if related to the recipient and/or the payer of the royalties, must have an equity of at least 50% of the expected annual amount of incoming royalties and prepay at least 50% of this equity amount to the so-called UBO without recourse, so the Dutch entity runs economical risk: if the royalties fall short of expectations, the entity starts to make losses that cannot be offset in any other way than against future Dutch profits (which will be rather low so a large part of the loss might well eventually evaporate). Traditional royalty conduit entities, owned by the client, have also become much trickier over the last three years because everybody is now focusing on ‘beneficial ownership’. A tax treaty requires a Dutch entity to meet two criteria before it can enjoy the reduced withholding tax rates of the countr(y)(ies) of the royalty payer(s). 1) A Dutch residency certificate, signed and stamped by the Dutch tax authorities. This was all that was looked at, usually, abroad upon a tax audit in the past. 2) But the tax treaties also demand that the Dutch entity receives the royalties for itself and not for a party outside the Netherlands (such as a BVI or Cayman Islands company). This aspect has been overlooked for a long time because it is difficult to apply. Most countries do not even have a definition of ‘beneficial ownership’ in their national laws so they have trouble defining the tax treaty term. For a tax inspector, in any country, basing an income adjustment or a tax rate adjustment on unclear grounds does not work and he knows it. OECD guidance to tax authorities on the meaning of ‘beneficial ownership’ In addition, the OECD is working hard on giving the tax inspectors of the world guidance on what is actually meant by ‘beneficial ownership’. When these final regulations will come out, this summer, foreign tax inspectors will for the first time have a real weapon to attack royalty conduit structures, based on new OECD commentary to the OECD Model Tax Treaty which has usually been the basis for the actual tax treaty between the Netherlands and the country of the royalty payers. The European Commission announced a GAAR provision in the Royalty and Interest Directive Another recent measure was the announcement by the European Commission to add a general anti abuse rule (GAAR) to the Interest and Royalty Directive. This new rule will have to be implemented in each EU country before 31/12/2015. It means that if a Dutch entity receives royalties from a related entity somewhere else in the EU, the tax inspector of that country may feel free to take the position that the Dutch entity has insufficient ‘substance’ under the laws and regulations of his country, so the reduction of the royalty withholding tax was given in error and the money will have to be paid back, including a fine and interest for late payment. And there is more to come... A further major threat are the so-called Limitation on Benefits (LOB) articles that are increasingly part of tax treaties: these articles basically deny any and all tax treaty benefits to a resident of the other country (the Netherlands) if the Dutch entity does not meet one or two ‘hard’ criteria either in the legal or in the substance area. Until 2012, the Netherlands had only one tax treaty with an LOB provision, with the USA. In 2012 Japan insisted on such an article for the new Dutch/Japanese tax treaty that went live per 1/1/2013 and got it, and in 2013 the Dutch government, unilaterally, promised to offer 23 other countries a treaty renegotiating, part of which would be the introduction of an LOB article...! How much risk are you prepared to take? So the picture for traditional Dutch royalty conduits is rapidly getting uncomfortable. The old ‘stability’ that these entities were known for, plus their low pricing, has been eliminated almost entirely and can only partially be regained by adjusting the set-up through adding considerable substance and obtaining expensive additional tax and legal input from the consultancy firms mentioned, that charge some €500/hour on average. And even then one cannot feel safe because foreign tax authorities may well ignore the Dutch new rules, saying they do not go far enough, and replace them by their own, much stricter, rules, like Germany and China have done. What are the basic steps to address the new risks and the cost associated therewith? It should be emphasized that both the old and new Dutch national ‘substance’ rules and the new European Commission’s instruction to countries to implement a GAAR in the European Interest and Royalty Directive, focus only on related companies. This has historically been so (in the EU case even by definition). These governmental solutions are therefore not applicable to our independent Dutch royalty conduit entities! So by trading your own Dutch royalty conduit BV for a similar entity, but now owned by Dutch residents, one breaks the chain and all of a sudden transfer pricing, benchmark studies, advance tax rulings etc. become obsolete. The simple way is to sell the Dutch conduit entity you currently employ to a Dutch provider, so all business features remain unaltered (especially the existing contracts with the royalty paying entities). The new price you will agree with the Dutch fiduciary service provider is a third party price by definition, not open to scrutiny by any tax authority. And the price setting is simple: the Dutch owner of the conduit entity will incur, for his own account, almost all expenses of setting up the Dutch structure and running it. The client just pays him a percentage of the royalties that flow through the providers entity. The only other expenses the client pays would be for: a) The contracts with the ultimate licensees (a cost the client would also incur if he owned the Dutch entity himself, because this contract is a business contract that can only be negotiated by him). The Dutch provider will of course also have a say about this contract because it will be one of his companies that will sign it. b) A second cost element that the client will have to pay for himself is the cost of setting up and running a website for the Dutch BV, as the expenses therefore are not under the control of the Dutch provider and could erode his gross margin to unacceptably low levels. There is no need for a Dutch royalty conduit to have its own website and the vast majority of them do not have one, but it is a possibility. c) The third and last exception to the rule that the Dutch owner/provider will pay for any and all costs of setting up and running the structure and all business risks of the entity used are his worry and not the client’s are the legal expenses associated with any business litigation by or against the Dutch company [eg. for patent or trade name infringements]. The Dutch entity is collecting royalties for his product or invention and the safekeeping of this invention or product is in the interest of the client, not in the interest of the Dutch intermediate, third party owned, legal entity per se. This rule again applies because litigation expenses can be very high and are outside the control of the Dutch provider. The margin the Dutch entity will make on the royalty flow through business is usually a declining percentage of the foreign royalty income after royalty withholding tax if any (in most cases, under the Dutch tax treaties, the foreign royalty tax will go down to zero). For example: • 6% on the first €500,000 of net after tax royalty income per annum; • 5% on the next €500,000 of net after tax royalty income per annum (ie. on income between 500K and 1,000K); • 4% on the next €1,000,000 of net after tax royalty income per annum (ie. on income between €1 million and €2 million); • - 3% on any further net after tax royalty income per annum (ie. on income exceeding €2 million) This then works out as follows: say that in a given year the Dutch subsidiary entity receives 1.8 million of net royalty income. Its reward, to be withheld from the onward payment of the royalties to the client, will then consist of: • 6% on 500K = 30K • Plus 5% on another 500K = 25K • Plus 4% on the residual 800K = 32K • Total margin for the Dutch entity: 30K +25K + 32K = 87K = 4.8% of total income. How to deal with LOB provisions? One element of an LOB provision is that the company that claims tax treaty benefits, can only do so if it has limited ‘types’ of owners (in fact: in the current Dutch tax treaties with the USA and Japan say they have to be Dutch owners or owners resident in another EU country). This element is of course also a threat to the traditional royalty conduit business because many royalty earners do not or no longer live in a high tax country or in the EU, so these LOB provisions already by themselves call for a change of ownership of ‘their’ Dutch royalty conduit entity to someone else (in line with the new Dutch rules and the new EU interest and royalty directive). But we solved this problem already in the above by having the Dutch royalty conduit company be owned by the Dutch service provider who also delivers the traditional services of hosting the Dutch entity and its management. But an LOB provision, also in cases where the Dutch conduit company is not owned by the client but by Dutch or EU residents, stipulates that a qualifying entity (the shares of which are held by the ‘proper’ owners) if it pays more than 50% of its incoming royalties out to non-qualifying persons in a tax deductible manner, the entity still disqualifies for treaty protection. This rule seems to kill most traditional royalty conduit companies in the longer run, because the Netherlands will increasingly allow LOB provisions in its tax treaties and, as mentioned, is nowadays even offering such LOB provisions to 23 countries with which it already has a tax treaty, but without an LOB provision. It is only a matter of time before the Netherlands will have 6, then 14, then 20 and perhaps in the future even 25 tax treaty partners with which it has agreed upon an LOB provision in the respective tax treaty. In my view it would be a bit foolish to wait till this happens and to then, all of a sudden, start to look for a Dutch royalty conduit structure that can withstand LOB provisions, rather than doing this now, at ease, when the threat is still ‘around the bend’. New structures will be needed to stay in line with the current and future tax developments In order to overcome this last but hard to circumvent new tax treaty development, we have, already for a few years, been using structures that consist of two Dutch entities: a Dutch parent BV and its wholly owned Subsidiary BV. Dividends from the subsidiary BV to its parent are non-deductible for tax purposes, so they meet the above-mentioned ‘50% test’: the entity which receives the royalties from abroad is not allowed to pay more than 50% of this income onwards unless this 50% is non-deductible. Dividends are non-tax deductible in the Netherlands, like almost anywhere else in the world. The double-BV structure or BV1/BV2 structure that we employ for clients easily meets both LOB criteria: 1) The owners of the Dutch royalty conduit are also the owners of its parent company; 2) And the onward payment at the level of the royalty receiving entity (‘BV2’) consists of a non-tax deductible dividend to its parent (BV1). See Exhibit 1 at the end of this article for a visual. This solution, without further planning, would seem to lead to a major Dutch tax problem: the subsidiary, if it cannot deduct the dividends it pays, will show a very high taxable profit which would undo all of the advantages of using the Netherlands as conduit jurisdiction. This can be resolved in two ways under Dutch national tax law, however. I will only elaborate on the least attractive of the two solution here, on which an advance tax ruling could even have been obtained in the past (not anymore, like I explained, under the revised Dutch substance rules from 1/1/2014 onwards). But one can do this also without an advance ruling, based on a thorough benchmark study and a transfer pricing report (needed because the two Dutch entities are related in a transfer pricing sense, even though they are unrelated to the ultimate recipient of the royalty income and also unrelated to the payers of the royalties).This briefly works as follows: the Dutch subsidiary might have obtained the right to collect the royalties from its 100% parent in lieu for a constant flow of interim dividends, for free. Under standard transfer pricing rules, this will then cause ‘mispricing’ and the conduit company can apply clear and constant case law from the Dutch supreme tax court that it can, for purposes of filing its corporate tax return, take a deduction for the royalties it should have paid if the parent had been a genuine third party. This tax deductible item would be close to the royalty amounts that the Dutch parent is paying to its client to obtain the exploitation license which gives rise to the royalty income in the first place... But there is even a better solution available in the Netherlands which resembles the US ‘check the box’ rules. Also the Dutch corporate tax system, on a very limited scale, contains elements whereby a Dutch subsidiary, upon election, becomes a ‘branch’ of its parent, which eliminates all transactions between the subsidiary and the parent, because one cannot sell a license to oneself, conclude a loan with oneself etc. Conclusions and recommendations In the above I have tried to explain that the Netherlands may still continue to be the best jurisdiction in the world to collect royalty type income, under its very favourable tax treaties with other countries, but that it would be naïve, when a new structure has to be set up, to do this based on old and outdated tax rules even if the traditional Dutch ‘trust sector’ does not seem to be aware of the three major attacks that have been launched over the last two years to stop what many, even in the Netherlands, see as ‘abuse’ of the Dutch tax treaties. And owners of existing traditional Dutch royalty conduit structures may want to consider how much risk they are taking by not changing their structures to make them ‘state of the art’. In exhibit 1 below, a basic diagram of our BV1/BV2 solution can be found. We suggest anyone who likes this idea to not start doing this at home, but to request our assistance in setting the new structure up and answering the list of questions that may well have resulted from reading the above carefully.


... the picture for traditional Dutch royalty conduits is rapidly getting uncomfortable