Withholding Tax1

Document Sample
Withholding Tax1 Powered By Docstoc
					Withholding taxes Professor Frederik Zimmer, University of Oslo, Norway 1. Introduction Since most of the discussions this day of the conference will be dedicated to taxation of wages and salaries, it is appropriate to point out that this section focuses on withholding taxes on investment income, and on dividends in particular. Withholding taxes on salaries and wages on the one hand and on dividends, interest and perhaps royalties on the other are very different things. In fact, what they have in common is the technical way of collecting the tax and not much else. The underlying policy issues are different and so are their relation to tax treaties and EU/EEA law.1 Withholding taxes on investment income are levied in the source state, normally on gross income at flat tax rates and as definitive taxes without any subsequent assessment, the maximum tax rate often being limited in tax treaties. Withholding taxes derive their name from the way they are collected: by being withheld by the debtor and paid by him directly to the government. From a domestic law point of view, withholding taxes (hereinafter wht/whts) on investment income are a technique to collect tax from non-resident taxpayers. From a tax treaty point of view, wht is first and foremost a technique by which the right to tax a certain item of income is allocated between the two states and normally divided between them. From the point of view of EU/EEA law, wht is primarily a problem: a possible discrimination or restriction contrary to the freedoms of establishment and movement of capital. Therefore, the ParentSubsidiary Directive and the Interest and Royalty Directive prohibit whts on direct investments but whts are still an issue within the EU for portfolio investments. It should be mentioned that these directives are not part of the EEA Agreement and therefore do not apply to Iceland, Liechtenstein and Norway (referred to as EFTA pillar states). However, Norway has never had whts on interest and royalty in domestic tax law. 2. Some general remarks on withholding taxes in domestic law and tax treaties Leaving the EU/EEA-related problems aside for a moment, which are the policy reasons for levying a wht – or put more generally: to tax non-residents on investment income? This leads to the basic issue of source vs. residence taxation, which is too big an issue to be dealt with comprehensively in this context. However, a few key issues should be mentioned: Should investment income be taxed in the state in which the capital is created or in the state in which it is used? Even the UN Model accepts some residence state taxation, but nevertheless the argument that investment income should be taxed only in the source state is still alive. The Andean Model of 1971 was based on that rule, and more recently, the general reporters at the Buenos Aires congress concluded that “exclusive or predominant taxation at source” should be favoured generally.2 They argue on the basis of inter-nation equity and – interestingly – point to the fact that the less progressive or even proportional tax rates on investment income reduce an important argument in favour of residence state taxation: that of taxing the total income of the taxpayer according to the ability to pay principle.

Except perhaps for artistes and sportsmen for whom the withholding tax is often the definitive tax in the state of source, and the same applies in some states for income from services as well, se section 3 below. 2 IFA (publ.): Cahiers de droit fiscal international, 2005, vol. 90a, p. 40 (general reporters: Professors Angel Schindel and Adolfo Atchabahian of Argentine). That statement does not apply exclusively to investment income.


So at least some degree of source state taxation of investment income is probably reasonable in treaties between developing and developed countries. However, between industrial states, as most of the European states, there is to a lesser extent a one way traffic of investments, so those arguments are not convincing to the same extent, even though some states certainly are net capital exporters and some importers. At the same time, there are certainly practical problems connected with the whts. Avoiding whts is the most important reason for treaty shopping, and treaty shopping, in turn, has led to complicated limitation on benefits clauses. Also, the beneficial owner concept is particularly important in a treaty shopping context. Accordingly, dropping whts would simplify the application of tax treaties considerably. The division of taxing rights embedded in most tax treaties for dividends and to some extent for interest and also royalties forces exemption states as well to rely on the credit method for such income. This leads to the well-known effect that low source state taxes favour the residence state rather than the taxpayer, which in turn raises the controversial issue of tax sparing and matching credits. It also sometimes occurs that the residence state is not prepared to grant a tax credit because it applies the exemption method.3 On the other hand, whts in a way secure taxation in cases where the residence state may have problems with being informed of the income (this technique is well known in the Savings Directive) and also when there are doubts as to which state is the residence state. In addition, the use of whts is an efficient way of collecting taxes levied on non-residents.4 From a purely domestic law perspective, levying whts on investment income may be sound, not least because that tax will in practice apply on payments to taxpayers resident in non treaty countries, among them the tax havens. It has also been argued that states are interested in having a high wht in order to induce other states to enter into treaty negotiations and improving their bargaining positions in such negotiations. However, this has to be weighted against the risk of losing investments from investors resident in non treaty countries. Thus states, after all, will probably be reluctant to introduce very high whts on investment income. One problem is that whts are levied on gross income. Even if there are normally no large costs connected with such investments – in particular portfolio investments – this nevertheless is a deviation from the principle of taxing net income. For instance, the taxpayer may have a loss on other sources of income in the state of source. Further, whts raise difficult issues regarding the procedural rights of taxpayers, as was illustrated in the Fokus Bank case: Normally, the taxpayer is not informed of the withholding and may have limited rights to appeal the case. Whts may be considered as a restriction under OECD Model Convention Art 24(5) because they are levied on foreign shareholders only. However, the commentaries, as probably amended 2008 (commentary 78), make clear that whts as such do not amount to a discrimination under this rule. There is reason to point out that the abolition of whts as such would not solve the issue of economic double or multiple taxation of company income; to reach that effect, either exemption or credit for underlying taxes is required, as illustrated by the Parent-Subsidiary Directive.
3 4

See as illustrations ECJ cases C-513/04 Kerckhaert Morres and C-170/05 Denkavit. As accepted by the ECJ in case C-290/04 Scorpio (regarding service income) paragraphs 35-37.


3. EU/EEA tax law: Withholding taxes acceptable at all? Turning to EC/EEA law, there are not only policy issues but also current legal issues to discuss. One basic question is whether whts on non-residents‟ investment income are acceptable at all under EU/EEA law, considering that residents are normally taxed the ordinary way by assessment for such income. To the extent whts may lead to higher taxes for non-residents than for residents, the issue is discussed below. Here, the focus is on the wht as such. Aspects which may indicate a restriction are the duty of the debtor to withhold the tax and pay it to the fisc, the liability following non-compliance and also that whts are often paid earlier than taxes levied on assessment. To the knowledge of this writer, this issue has not been dealt with by the ECJ in cases concerning investment income.5 However, it was dealt with in cases C-290/04 Scorpio and C443/04 Commission v. Belgium concerning income from services. In both cases, according to domestic law, the recipients of services from non-resident service providers were obliged to withhold part of the consideration at source and would incur a liability if that were not done. In both cases, the ECJ found that the withholding obligation and the liability following the non-performance of this obligation would deter companies from engaging non-resident service-providers and therefore amounted to an obstacle prohibited by the freedom of services rule. However, in Scorpio such legislation was found to be justified “by the need to ensure the effective collection of income tax” and was further regarded as a proportionate means for that purpose (paragraphs 35 and 37). The potential liability was justified for the same reasons (paragraph 38). In Commission v. Belgium however, the Court found no justification. This somewhat puzzling difference may perhaps be explained by the fact that the withholding obligation seems to cover somewhat different realities in the two cases. In Scorpio, the withholding tax reflected a real tax obligation for the non-resident service provider whereas in Commission v. Belgium the non-resident service provider would often have no tax obligation in Belgium and, therefore, the withholding obligation to a larger extent seems to have had the purpose of combating tax fraud.6 In this respect, wht on investment income is more similar to Scorpio than to Commission v. Belgium. However, it has been pointed out by distinguished writers that Scorpio concerned a case in which the Mutual Assistance Directive or similar rules did not apply and that the Court put some emphasis on that fact (paragraph 36).7 The conclusion in case this directive or similar rules apply is therefore somewhat uncertain.


In cases C-170/05 Denkavit and C-379/05 Amurta the different treatment of resident and non-resident shareholders was certainly in focus but not the fact that the tax on non-residents was levied as a withholding tax. In Fokus Bank, the issue was raised during the oral hearing whether the liability following non-compliance of the duty to withhold taxes on distributed dividends in itself amounted to a restriction. However, as the referring court had not asked that question it was not discussed by the EFTA Court (paragraphs 46 and 47). 6 The fact that Scorpio was decided in Grand Chamber and Commission v. Belgium a chamber of only three judges five weeks after Scorpio and without any reference to that decision indicates that the Court must have considered the cases to deal with rather different situations. 7 See Daniel Garabedian and Jacques Malherbe: Cross-Border Dividend Taxation: Testing the Belgian Rules against the ECJ Case Law (or Testing the ECJ Case Law against the Belgian Rules), in Hinnekens/Hinnekens (eds.): A Vision of Taxes within and outside European Borders. Festschrift in honor of Prof. Dr. Frans Vanistendael, 2008, p. 397 at p. 407, and Guglielmo Maisto: The European Court of Justice and Domestic Law of Tax Procedure: A Critical Analysis, in the same volume p. 635 at p. 638.


A further issue is whether it is acceptable that taxes are withheld at the time of payment but subject to later recalculation in order to get the taxation of non-residents in line with that of residents. An example is the Norwegian rule on dividends paid to individuals resident in other EEA States. Resident personal shareholders are entitled to a so-called protective allowance which reflects the ordinary yield of their investment. This allowance is not taken into consideration when taxes on dividends to non-resident shareholders are withheld. However, shareholders resident in other EEA States than Norway are entitled to a recalculation later in which this allowance is taken into account and the shareholders are also entitled to interest on the refunded amount. The reason for this technique is that the entitlement to the allowance is allocated to the person who owns the share at year end and also that the amount of the protective allowance varies from person to person. 4. The different treatment of resident and non-resident shareholders: Fokus Bank, Denkavit, Amurta In practice, the question has been put to its extremes in imputations systems, in which imputation credits were granted to resident shareholders receiving dividends from resident companies whereas a withholding tax was levied on dividends distributed to non-resident shareholders without any imputation credit being granted to them. The question whether such rules amount to a restriction/discrimination conflicting with the rules on freedom of establishment and free movement of capital has been an issue in several recent cases. Three cases stand out regarding this question: The EFTA Court8 case E-1/04 Fokus Bank, and the ECJ cases C-170/05 Denkavit and case C-379/05 Amurta. In all the three cases, the courts found that the different treatment amounted to a restriction contrary to the rules on free movement of capital or the freedom of establishment. In the context of the taxation of distributed dividends, resident and non-resident shareholders are in objectively comparable situations when taxes are levied on non-residents. As the ECJ states in Denkavit (paragraph. 36):
“In the present case, parent companies receiving dividends paid by resident subsidiaries, are, as regards the taxation in France of those dividends, in a comparable situation, whether they receive those dividends as resident parent companies or as non-resident parent companies which have a fixed place of business in France, or as non-resident parent companies which do not have a fixed place of business in France.”

Turning to the significance of the tax treatment in the other state and the possible effect of the tax treaty, it becomes more difficult. The question is whether a tax credit in the other state (the residence state of the shareholder) can save the right of the residence state of the distributing company to levy the wht. (Whether this is regarded as part of the restriction discussion or as a justification varies; in Fokus Bank it is seen as under the justification perspective, in Denkavit probably under restriction).

The EFTA Court rules on the interpretation of the EEA Agreement as applied in the EFTA pillar states of the EEA (Iceland, Liechtenstein and Norway). The Court has three judges and no Advocate General. The EFTA Court tends to follow the reasoning of the ECJ very closely, which is good because otherwise we might have ended up with different interpretations of the EEA Agreement depending on whether it is interpreted by the ECJ or the EFTA Court; the ECJ has made clear that the principles of interpretation are the same regarding the EEA Agreement and the EC Treaty, see for instance case C-104/06 Commission v. Sweden paragraphs 31-33. However, the situation is not always identical, in particular because the Mutual Assistance Directive does not apply in EEA, and that may lead to differences in reasoning and to different conclusions.


In Fokus Bank – which was the first of these cases to be decided (november 2004) – the EFTA Court is completely negative:
“A Contracting Party cannot shift its obligation to comply with the EEA Agreement to another Contracting Party by relying on the latter to make good for discrimination and disadvantages caused by the former‟s legislation” (paragraph 37).

It did not help that the right to the tax credit is based on the tax treaty:
“… permitting derogation from the fundamental principle of free movement of capital laid down in Article 40 EEA on the ground of safeguarding the cohesion of the international tax system would amount to giving bilateral tax agreements preference over EEA law” (paragraph 31).

One may wonder whether this negative view is due to the fact that the issue was raised in the context of a justification based on the cohesion of the tax system and not as part of the question whether a restriction existed in the first place. Regarded as part of the question of whether a restriction exists, it may look otherwise. Anyway, in Denkavit the ECJ proclaimed the opposite view: The tax treaty was regarded as “part of the legal framework” which the Court “must take into account in order to provide an interpretation of Community law that is relevant to the national court” (paragraph 45) and the same is reiterated in Amurta (paragraph 80). However, as we know, this did not help France in Denkavit, as its wht was not credited in the Netherlands because of the exemption system applied there. Amurta brings us one step closer to an answer. In this case, the shareholder was resident in Portugal and owned shares in a Dutch company. Portugal obviously had a domestic credit rule. The tax treaty in question is based on the principle of ordinary credit. The Court first takes the view that domestic Portuguese credit rules would not help, because the Netherlands “is under a duty to ensure that, … recipient companies established in another Member State are subject to the same treatment as recipient companies established in the Netherlands” (paragraph 77). Then turning to the significance of the tax treaty, which becomes all important, the judgment abruptly stops by stating that the national court had not referred that question to the Court.9 Therefore, the national court by itself will have to determine “whether account should be taken … of the DTC” (paragraph 83) and if so “whether it enables the effects of the restriction on the free movement of capital to be neutralised” (paragraph 84). In the opinion of this writer, that must mean that there is no restriction if the Dutch tax is fully credited in Portugal. If this is not so, one would have expected the Court to say that clearly. Also, the reference in paragraph 83 to paragraph 28, where the Court states that there is a restriction if dividends distributed to non-resident shareholders are taxed less favourably than dividends distributed to

This conclusion seems questionable. The referring court asked as its second question whether the answer to the first question, on whether there was restriction, would depend on ”whether the State of residence of the foreign shareholder/company … grants that shareholder/company full credit for Netherlands dividend tax” (AG paragraph 17). This question is by its wording not limited to credit according to domestic law and the use of the term “full credit” need not be used as excluding credit under the treaty even if that treaty is based on the principle of ordinary credit. The reasoning behind the question of the referring court may perfectly well have included a situation where the Dutch tax was fully credited in Portugal under the tax treaty.


Dutch shareholders because that would deter foreign companies from investing in the Netherlands, points in the same direction. If the Dutch tax is fully credited in Portugal, then there will be no such unfavourable treatment which would deter Portuguese companies from investing in the Netherlands. Lastly, the Court very clearly keeps that option open when stating (in paragraph 79):
“However, it cannot be excluded that a Member State may succeed in ensuring compliance with its obligations under the Treaty through the conclusion of a convention for the avoidance of double taxation.”

However, the remarks from the Advocate General in that case may raise doubts. In paragraph 87 and 88 of his opinion he seems to express that only “full credit” as opposed to “ordinary credit” would fully mitigate the discrimination caused by the wht. He equates ordinary credit with “partial tax credit” and states that such a rule would never fully eliminate the effects of the wht in the source country. Adding to the confusion, in a footnote (no 42) he adds that full mitigation would in fact take place “only if the same tax rate were applied in the Netherlands and Portugal” (italics added), obviously excluding cases where the tax rates in Portugal exceed those of the Netherlands and therefore allow for a full credit of the Dutch taxes. Following this line of reasoning, the credit in the home state of the shareholder would only rarely help to mitigate the restriction caused by the wht. These remarks of the Advocate General seem confusing and not convincing to this writer. Reactions to Amurta have been rather diverse. At the one end of the spectrum, Peter Wattel criticises the Court for not accepting credit under domestic law as sufficient, arguing that the taxpayer would not care whether the credit is based on domestic law or on a treaty and that domestic law provides the same degree of certainty as treaty law. As far as the effect of the credit under the treaty is concerned he predicts that it suffices “to the extent in which the home State credit absorbs the discriminatory source State taxation”.10 At the other end of the spectrum, Lee Sheppard writes that the Court, contrary to the Advocate General, “refused to rely on the existence of a treaty for compliance with the EC Treaty”.11 This may refer to the fact that the Court did not conclude on the matter but may also be read as if the Court did not accept the relevance of the treaty at all, which is difficult to harmonise with what the Court in fact writes, both in Amurta and in Denkavit. Michael Lang seems to take the statement of the Advocate General quite literally and require a full tax credit:
“The Court uses the term „full tax credit‟, which obviously has to be distinguished from the „ordinary credit‟. … the interpretation issue seems to be quite obvious and the existence of a „full tax credit‟ very hypothetical”.12

He must be referring to a full tax credit in the tax treaty meaning, as opposed to ordinary credit.

10 11

Ben J.M. Terra and Peter J. Wattel: European Tax Law. Fifth edition. 2008, p. 744-45. Tax Notes International, 31. Dec. 2007, p. 1346. 12 Michael Lang: ECJ case law on cross-border dividend taxation – recent developments, in Intertax 2/2008 p, 67 at p. 71.


The various writers‟ views on this question depend to a large extent on their views more generally of whether the tax situation in the other state should be taken into account or not, often referred to as the overall (or internal market) and per-country approach respectively. Michael Lang together with for instance Dennis Weber, belongs to the per-country school,13 whereas for instance Peter Wattel and in particular Eric Kemmeren belong to the overall school.14 Similarly, Fokus Bank seems to rely on a per-country approach whereas Denkavit and Amurta imply an overall approach. It should be added that it seems to have been unclear during the proceeding which rules in fact apply in Portugal. (Nothing is yet known as to the further handling by the national Dutch court.) One may ask what the legal situation is in this regard in the EFTA pillar of the EEA in cases where the wht is fully credited in the resident country of the shareholder. Is the reasoning of the EFTA Court – excluding the tax treaties as a matter of principle – still valid, or will the ECJ reasoning – as understood above – prevail? This is a question of some practical importance because the Fokus Bank case was only one of several similar cases and some shareholders obviously were residents of credit countries. Imputation systems are all but history in Europe, to a large extent because of their problematic relationship to EU/EEA law. In classical or semi-classical systems, which have instead been introduced in several states, also dividends to resident shareholders are taxed effectively, and a different treatment between the taxation of resident and non-resident shareholders is therefore not as obvious. However, the fact that the wht is levied on the gross income and that different tax rates often apply may raise similar problems as under the imputation systems. Depending on the circumstances in each case, the tax on dividends received by non-resident shareholders may be higher than on dividends received by resident shareholders. According to the ECJ case law it must be decided on a case by case basis by the national court whether or not non-resident shareholders are treated less favourably than resident shareholders.15 So even


Michael Lang in Intertax 2/2008 p. 72 (arguing that the overall position leads the Court into the position of the legislator and that the per-ountry approach “increases the pressure on the Community legislator to take action or on the Member States to coordinate their tax rules voluntarily”) and Dennis Weber: In search of a (new) equilibrium between tax sovereignty and the freedom of movement within the EC (2006). 14 See Terra and Wattel op. cit. footnote 10, p. 741, in particular p. 744, and Eric C.C.M. Kemmeren: The Internal Market Approach Should Prevail over the Single Country Approach, in: Hinnekens/Hinnekens (eds.) op. cit. footnote 7, p. 555. See also Gerhard T.K. Meussen: Denkavit Internationaal: The Practical Issues, in European Taxation 2007 p. 244 at p. 245. 15 See case C-265/04 Bouanich: Redemption of shares was according to Swedish rules taxed as capital gain for resident shareholders but as dividend for non-resident shareholders. For residents the cost price of the shares should be deducted from the sales price and the gain taxed at a rate of 30 %. Non-residents had the right (according to tax treaty) to deduct the nominal value of the shares and the net amount should be taxed at a rate of 15 % (again according to the tax treaty). The Court concluded: “55. It is therefore a matter for the national court to determine in the proceedings before it whether the fact that non-resident shareholders are permitted to deduct the nominal value and are liable to a maximum tax rate of 15% amounts to treatment that is no less favourable than that afforded to resident shareholders, who have the right to deduct the cost of acquisition and are taxed at a rate of 30%”. For a similar approach concerning wht on income from services, see cases C-234/01 Gerritse (paragraph 54 in particular) and C-290/04 Scorpio. Daniel Garabedian and Jacques Malherbe, op. cit. footnote 7, p. 405 suggest that if the taxpayer has been able to claim a deduction for costs in his state of residence, then wht based on the gross amount in the source state and leading to higher taxes there than for resident shareholders of that state, should not be considered contrary to EU/EEA rules.


if imputation systems are abandoned the issue of possibly discriminating whts is still here.

5. The treatment of the withholding taxes in the state of residence of the shareholder: Kerckhaert Morres Whts are normally creditable in the residence state of the creditor if that state taxes the income in question. This is not always the case, however, and the question arises as to whether EU/EEA law requires the residence state to grant such a credit (or provide other kinds of tax relief) even if domestic law in the residence state does not treat foreign source income less favourably than domestic source income. This was the issue in case C-513/04 Kerckhaert Morres. The taxpayers were residents of Belgium and received dividends from a French company. That dividend had been taxed in France by way of a wht. It was also taxed in Belgium, and credit for the French wht was denied. Belgium taxed both foreign source and domestic source dividends according to the same rules and tax rates (25 %). The ECJ found no breach of the EC Treaty “as the Belgian tax legislation does not make any distinction between dividends from companies established in Belgium and dividends from companies established in another Member State” (paragraph 17). This distinguished the case from the Manninen case (C-319/02) and other cases, in which foreign source dividends were taxed less favourably than domestic source dividends under domestic law of that state. Comparing Kerckhaert Morres with Denkavit, one may wonder why the taxpayers did not challenge the French wht instead. An answer is perhaps found in paragraph 25-26 of the opinion of the Advocate General: It seems as if the taxpayers in fact received a tax credit in France and the Advocate General concludes that they were “not worse off in comparison to those receiving Belgian-source dividends”. The judgment of the ECJ does not explain why the Court did not decide the case on that basis but chose a more controversial argument. The explanation is perhaps that the credit was based on the tax treaty whereas the question from the national court did not make a reference to the tax treaty.16 Further, regardless of France‟s duties under the tax treaty, that state would, under the principle of Denkavit, have an obligation under EC law to grant the same relief to non-resident shareholders as to resident shareholders, as is pointed out by the Advocate General in Kerckhaert Morres (opinion paragraph 39).17 It is clear, however, that a heavier taxation of cross-border dividends may result from this state of affairs if the Denkavit rule is not applicable in the residence state of the distributing company (source state) because that state does not discriminate non-resident shareholders. Assume that the tax rate on dividends in both the source state (S) and the residence state (R) of the shareholder is 20 % and that these taxes are levied regardless of whether the shareholder is resident or non-resident in the source state (in the latter case the dividend tax is levied as a wht) and regardless of whether it is received from resident or non-resident companies in the residence state of the shareholder. In this case, dividends from a company in both S and R to resident shareholders in these countries respectively are taxed at 20 % but

See Lieven A. Denys: The ECJ Case Law on Cross-Border Dividends Revisited, in European Taxation 2007 p. 221 at p. 228 and 232. It is understood that the question of whether the taxpayers were entitled to a tax credit in Belgium under the Belgium-France tax treaty is still pending, see Daniel Garabedian and Jacques Malherbe op. cit. footnote 7, p. 419 with footnote 64. 17 See also Lieven A. Denys op. cit. footnote 16 at p. 228.


dividends distributed from a company in S to a shareholder in R are taxed at 20 % both in S and R. For instance if the dividend is 100, then the tax is 20 in R and again 20 in R, altogether 40. According to the case law, this is no infringement of EU/EEA rules but “adverse consequences which … result from the exercise in parallel by two Member States of their fiscal sovereignty” (Kerckhaert Morres paragraph 20).

6. Inbound and outbound imputation credits: Manninen, Fokus Bank and ACT Group It has sometimes been suggested that the Fokus Bank (and Denkavit) and Manninen cases, seen together may lead to (unjustified) double tax credits.18 Also, a possible disharmony between Fokus Bank (and Denkavit) and ACT Group has been suggested.19 However, Fokus Bank must be understood in the context of the Norwegian rules applicable at that time. Norway certainly had an imputation system but excess credits were not refundable, neither for resident nor for non-resident shareholders. Thus, the issue in Fokus Bank was restricted to the question of the right of the non-resident shareholders to a tax credit to be deducted from withholding taxes calculated in Norway. The question of a refund for credits in excess of the tax liability in Norway, which was the issue in ACT Group, did not arise because there was obviously no discrimination on this point since not even resident shareholders were granted a refund. Therefore, nothing in Fokus Bank suggests that the residence state of the distributing company will have to surrender company taxes to the residence states of the shareholders. And in ACT Group, the ECJ found that non-resident shareholders who were not taxed in the residence state of the distributing company were not in the same position as a shareholder resident in that country. Therefore, there is no contradiction between Fokus Bank (and Denkavit) and ACT Group.20 Further, considering this scope of Fokus Bank there is no difficulty in reconciling it with Manninen: Fokus Bank requires the residence state of the distributing company to reduce or eliminate the wht in order to reach equal treatment of resident and non-resident shareholders in the residence state of the distributing company; that credit does not reduce the company tax

See Peter Wattel in Ben J.M. Terra and Peter J. Wattel: European Tax Law, Fourth edition, 2005, p. 128: “The result is thus that … the UK shareholders get two imputation credits (one in their home State on the basis of Manninen and one from the source State on the basis of Fokus Bank), which means that they get more credit than they paid tax for,…”. However, he is more understanding in the fifth edition of the book at p. 188. See also Michael J. Graetz and Alvin C. Warren Jr.: Dividend Taxation in Europe: When the ECJ makes Tax Policy, in Common Market Law Review 44 (2007) p. 1577 at p. 1600: “…Manninen and Fokus might be read together to require both source and residence countries to grant imputation credits to cross-border dividends, which could make the tax burden on international investment less than that on domestic investment.”; see also p. 1603 (“the EFTA Court read the ECJ decisions to require …” imputation credits to both inbound and outbound dividends), p. 1611 footnote 109 (“the EFTA Court‟s analysis in Fokus does not turn on the distinction” between “imposition of a withholding tax and failure to extend imputation credits on outgoing dividends to individual shareholders”, 19 See Graetz and Warren op. cit. footnote 18 at p. 1616 (“The EFTA Court‟s holding in Fokus arguably requires such an extension…”, that is extension of imputation credits to non-resident shareholders which was denied in ACT Group). See also Peter Wattel op. cit. footnote 18 fifth edition at p. 188 (a contradiction “at first sight” but “the two Courts seem to agree after all”). Similarly Frans Vanistendael: Denkavit Internationaal: The Balance between Fiscal Sovereignty and the Fundamental Freedoms?, in European Taxation 2007 p. 210 at p. 211-212 (“Prima facie, the decision in Denkavit Internationaal seems to be in contradiction with the decision in Test Claimants Class IV…”). 20 See Wattel and Vanistendael op. cit. footnote 19, Thierry Pons: The Denkavit Internatioinaal Case and Its Consequences: The Limit between Distortion and Discrimination?, in European Taxation 2007 p. 216. Lieven A Denys op. cit. footnote 17 at p. 230, and Suzanne Kingston: A Light in the Darkness: Recent Developments in the ECJ‟s Direct Tax Jurisprudence, in Common Market Law Review 2007 p. 1321 at p. 1343.


in that state. Manninen requires the residence state of the shareholders to provide a credit for company taxes paid by non-resident companies if that state grants such a credit to shareholders in resident companies. In fact, both credits are necessary if cross-border dividends shall be treated no less favourably than domestic dividends. This can be shown by a simple example. Assume that in both state R and state S the company tax rate is 30 % and that resident shareholders in both states are granted an imputation credit (or are exempt from tax); assume further that state R levies a 10 % wht on dividends paid to non-resident shareholders. If the income of a company in state S is 100, then company tax in state R is 30, and if the remaining 70 is distributed a wht of 7 is triggered whereas the tax on the dividend in state R is 21 (30 % of 70) if no credit or exemption is granted. According to Fokus Bank, state S must, however, reduce the wht levied in that state to zero; according to Manninen state R must grant an imputation credit which reduces the tax on the shareholders to zero (because the tax rates are the same in both countries) or exempt the dividends. The result is that shareholders in the state S company who are resident in state R are treated in the same way as shareholders resident in state R which are shareholders in a state R resident company, and that they are treated in the same way in state S as state S resident shareholders in a state S resident company. In both cases, the only remaining tax is the 30 % company tax in state S. 7. Withholding taxes and taxpayers’ rights The Fokus Bank case also raises and illustrates some issues concerning taxpayer‟s rights which do not seem to have been dealt with in the other cases.21 Technically, the tax is collected by the distributing company as a wht. In general, the tax authorities deal with the distributing company only and not with the shareholders who are the real taxpayers. The taxpayers are normally not informed of the withholding but they can normally see from the information they get from the distributing company that a tax has been withheld. Further, they are normally not notified of a possible re-assessment (as in Fokus Bank) and such re-assessment cannot as easily be seen from information from the company. Whether the taxpayers in Fokus Bank had a right to appeal was doubtful at that time. In Fokus Bank, these issues were put to their extremes because this case started out as a tax avoidance, dividend stripping case. The non-resident shareholders in Fokus Bank had “parked” their shares in Norwegian institutions for a few days around the time of distribution of the dividend. That was not accepted by the tax authorities who regarded the non-residents as the real shareholders. However, the dividend, of course, had already been distributed to the formal shareholders without any wht deducted (since the formal shareholders were residents of Norway and therefore not taxable). They did not get any information of the re-assessment and therefore had no practical opportunity to appeal, if such a right existed at all. That state of affairs was not accepted by the EFTA Court which stated that the rule on freedom of movement of capital, while primarily dealing with substantial issues, requires equal treatment “also with regard to procedural rights in so far as procedural rights are a prerequisite to the protection of substantive rights under the EEA Agreement” (paragraph 43). The taxpayers could have argued, for instance, that the treatment was contrary to the EEA Agreement or that no tax avoidance scheme was at hand in the first place.

See Guglielmo Maisto op. cit. footnote 7 p. 635.


The Norwegian state also argued that such systems were common in other states as well but the Court dismissed that as irrelevant (paragraph 44). The Norwegian procedural statute rules have not been changed since the judgment in Fokus Bank but the administrative practice has: A right to appeal the assessment is now granted in practice, and non-resident shareholders shall now be informed of reassessments of the wht. However, there is still no right to be informed of the regular levying of the wht and the last word has probably not been said.

8. Concluding remarks Whts on investment income creates considerable difficulties not only under EU/EEA law but also under tax treaty law and domestic law. The recent case law of the ECJ (and the EFTA Court) may be understood to indicate that the days of whts within EU/EEA are counted (also outside the scope of the Parent-Subsidiary Directive). However, this case law (including case law regarding inbound dividends) did not lead to the abolition of whts but instead to the abolition of imputation systems. Under the new regimes – many of them of a semi-classical type – the discrimination of cross-border dividends is much less common. So, perhaps the whts will survive after all.