En privat sajt med många skattetips och en guldgruva för dem som arbetar aktivt med skatter!

Denna webbplats administreras av Skattepunkten AB
Skattepunkten AB ger Dig professionell skatterådgivning

Kontakta Skattepunkten AB när Du behöver professionell hjälp av en skattekonsult.

Webb-journal on international taxation (Sweden) no 2/2005
by Peter Sundgren
September 2005


Double tax conventions and double non-taxation

I have a friend whose tax situation is so simple that he files his tax return by short message service on his cell phone. But he always complains – which is no surprise as far as Swedish tax payers are concerned - about the amount of taxes that he keeps paying. When I informed him that certain people deriving foreign income legally can avoid taxation altogether by using international treaties for the avoidance of double taxation, his reaction changed from surprise to jealousy and then to sheer indignation. “But it is absurd ” he protested, “that an agreement the very purpose of which is to relieve double, and I repeat, double taxation, winds up doing away with taxation altogether! Something must be very wrong! And, to make matters worse, all this is effected under carefully conducted negotiations between governmental tax experts and then approved by parliaments who in normal cases demonstrate an all too ardent desire to tax us other poor devils!” 
And surely one must have understanding for his reactions!
For the readers of this journal it is of course a well-known fact that double non-taxation may occur quite frequently under double tax treaties. In my opinion, which, I would suggest, is shared by all responsible analysts, it is a very serious problem that treaties for the avoidance of double taxation are in fact exploited as instruments for escaping tax altogether in both of the contracting states. Everybody will agree that double non-taxation, apart from such cases where tax sparing rules etc. provide for quite deliberate double reliefs, is unintended and in conflict with the purposes and objectives of tax treaties. Double non-taxation  thus serves to undermine the integrity of the treaties and it is incompatible with the principle of a fair sharing of the tax burden. Double non-taxation erodes the respect of ordinary taxpayers of the tax system as a whole and is harmful to tax compliance.
So what has been done to counteract double non-taxation?  The answer is  - very little. There are indeed subject-to-tax rules in certain treaties that may take care of some of the problems but they are relatively few and have, as far as Sweden is concerned, frequently proved to be too narrowly construed. The Swedish government has also taken a somewhat lethargic attitude towards effecting changes of tax treaties where double non-taxation has emerged, probably because it is such a time consuming business. (Only after about six years of negotiation and finally a termination of the treaty was it possible for Sweden to close the double non-taxation loophole in the treaty with Belgium reported below.)
Nor has scientific research in matters regarding double non-taxation been very comprehensive. Until  quite recently. Thus in June 2002 a High Level Scientific Conference was held at the university of Vienna on the “Avoidance of Double non-taxation” (Swedish national reporter Eva Fransberg, Stockholm School of Economics) and, most importantly, the 2004 Congress of the International Fiscal Association (IFA) in Vienna also chose double non-taxation as one of its main subjects. See Cahiers de Droit Fiscal International 2004 Volume 89a. (Swedish national reporter, Maria Hilling, LL.D. Candidate, Jönköping International Business School.)


How big is the problem?

The national reports to the IFA congress give no indication of the extent of the problem as such. The very fact, however, that IFA has arranged a special congress to discuss the issue gives an indication that there is serious concern about the matter. In Sweden there are no statistics to prove the scale of double non-taxation but it is nevertheless clear that in the recent past it has given rise to tax base losses amounting to billions of Swedish kronas. Thus, in the beginning of the 1980-s, and in the wake of a boom on the Swedish stock market, an exodus took place of Swedes moving to the UK when it was found that the treaty with that country allowed for a double non-taxation of un-remitted capital gains on disposals of Swedish shares. The amount of tax litigated in just one specific case regarding the heiress to one of Sweden’s greatest fortunes, was, at the time, the highest ever as far as an individual taxpayer was concerned. In the so called Bourghardt case that gave rise to the advance ruling in the matter, the municipal (capital gains) tax projected in 1985 by the small municipality,  which had been the domicile of Bourghardt and who moved to the UK to sell Swedish shares,  was so substantial that it considerably reduced the municipal equalization tax amount to be collected from the federal treasury. This, together with the eventual shortfall of capital gains tax under the Supreme Court decision of the subject to remittance rule in the Sweden-UK treaty, was so substantial that the municipality had to increase its tax rate for all its residents in the following year!
Another double non-taxation situation which gave rise to a gigantic erosion of the Swedish tax base, appeared in Sweden’s former treaty with Belgium, in which Sweden, quite surprisingly, had waived its right as source state to tax capital gains on the alienation of Swedish situs immovable property, where at the same time Belgium did not tax the gains on foreign real property.[1] In the late eighties, when the real estate market exploded, Swedish owners of immovable property who wished to exploit this double non-taxation situation therefore moved to Belgium to such a degree that it registered an all time high in the official statistics of emigrants to Belgium! And then, a couple of years later, when the tax exempt gains had been realized and the sellers had resided in Belgium “long enough”, the number of people returning back to Sweden also rose sharply.
Another well known tax ploy to achieve double non-taxation, albeit of a slightly different nature, is for Swedish retirees moving to France to transform life long pension schemes into short term benefits which are tax exempt in Sweden (again as source state) under the treaty, and tax free in France under French domestic tax law. When Sweden, during the negotiations with France for a new treaty, tried to gain primary taxing rights to Swedish source pensions, an influential lobby of Swedish pensioners in France even managed to inspire a member of the French parliament, during ongoing treaty negotiations, to give a passionate speech in the National Assembly defending a status quo of the Sweden-France tax treaty. And still, France remains a tax paradise for Swedish retirees.
Swedish IFA branch involvement.
Being an IFA member myself I had great expectations when the topic of double non-taxation was chosen for the 2004 Vienna congress. However, my misgivings about the treatment hereof emerged already with the presentation of the directives. The scope of the subject was thus narrowed to a rather one-dimensional academic and technical level mostly concerned with, admittedly, the intellectually stimulating but rather abstract problem of tax treaty interpretation in general and of the OECD-model in particular, focusing on whether or not or to what extent the purpose of treaties is to prevent double non-taxation. And frankly, the answer to this question, even if it were positive, is still not of very great practical consequence. Tax treaties have been around now for almost a century and, as can be gleaned from what has been mentioned above, double non-taxation still exists and it is growing into a big problem requiring more determined counteracting measures. Indeed, in his ‘bible’  “On Double Taxation Conventions”, Klaus Vogel even suggests that double non-taxation, not being subject to any competent agreement procedure, is even more serious than double taxation.  
In the autumn of 2003 the Swedish IFA branch, which it always does, arranged a meeting where the national reporters are given an opportunity to present a draft of their reports and to discuss them with the participating members and, which is the main purpose, to benefit from whatever ideas and suggestions that might be put forward. I had numerous points and suggestions to bring up regarding this topic. So many indeed that I even took the trouble of submitting a written memorandum thereof, which was distributed to the participants well in advance of the meeting. This, however, gave rise to a very sharp reaction from some fellow members (Jan Francke, Leif Mutén, Sven-Olof Lodin and Roger Persson-Österman) who filed an open letter to the meeting - at which none of them was present - rejecting my suggestions as improper and destructive.  The letter contained nothing of substance regarding double non-taxation but just recommended that the draft report, which incidentally had already been scrutinized and ‘approved’ by Messrs. Francke and Mutén  (see page 655 of the Cahiers), be adopted without further ado. (Some of my suggestions were however taken into consideration in the final version of the report.)
I duly responded to the letter.[2]
Scope of the IFA congress reports.
My main disappointment regarding the reports to the congress stems not so much from what is in them but rather from what is missing therein.
Most importantly the reports of the Cahiers fail to focus on the socio-economic aspects of double non-taxation and to imprint upon tax legislators the urgency of the situation. Nor is there in the reports any deeper empirical investigation of the extent of double non-taxation in the countries represented. No attempt has been made to determine the specific situations where double non-taxation occurs or how it emerges.  Certain types of double non-taxation that follow for instance from double deductions of expense have not been explored. Other questions that could have been asked are: What types of income are involved? Which treaties have been exploited? What was done about it? Does double non-taxation occur under domestic law provisions? 
Lacking more or less completely from both the national and general reports are any solutions or suggestions for counteracting double non-taxation. All that comes out of the general report is that the OECD and treaty negotiators don’t seem to have been focused enough on the problem and that the Model Treaty and its Commentaries suffer from a lack of balance. The very first sentence of the general report that “in recent years international tax planning has been subject to continuous improvement where tax payers and their advisers are benefiting from taxation loopholes resulting from increasing situations of double non-taxation”, does not signal a very great concern about the gravity of the problem. Rather it seems as if the general reporter, Professor Michael Lang (Austria), is more disturbed by the fact that “tax authorities are seeking to find new ways of interpretation that will contribute to ensuring that tax is levied at least once, i.e. that lead to single taxation of the income” (second paragraph of the introduction of the general report).
In the late autumn of 2003 I volunteered for a seat on the discussion panel of the congress for this subject. This however was declined because the panels were already complete. Shortly before the congress, which, after all, I was unable to attend, I addressed a paper to the discussion panel covering my views on this subject. After the congress I received an answer hereto from professor Lang.[3]
The following presentation will embrace the main points of  the memorandums I prepared for the branch meeting and the congress proceedings and also includes some comments to certain of the points discussed at the congress according to the summary of the congress discussions prepared by the International Bureau of Fiscal Documentation and a report of the congress in the Swedish tax periodical Svensk Skattetidning (9/2004)

Is the avoidance of double non-taxation the object and purpose of a tax treaty?

This very first question addressed in the general report is somewhat puzzling because it suggests that, already in a non-treaty situation, there is a state of double non-taxation that must be cured by the treaty. This, as everyone recognizes, is of course not the case. On the contrary it is the treaty itself, which in its interaction with domestic law in one of the treaty making states may give rise to double non-taxation. Maybe, therefore, it would have been more appropriate to ask to what extent it has been the object and purpose of treaty making states to prevent or allow double non-taxation in their treaties. As will be further explored below double non-taxation in all but very few cases is an undesired and almost always unintentional effect. In this light it would seem natural for the interpreter to attempt to counteract such unintentional  effects.
But the general conclusion of the reports seems to suggest that the prevention of double non-taxation is not a primary concern of tax treaties. The Italian national reporters Paolo Conci and Siegfried Mayr thus conclude for instance that “the prevention of double non-taxation may not, in our view, be regarded as the aim of the treaties” but, on the other hand, they also admit that “double non-taxation may also not be regarded as the aim of treaties.” Professor Lang, in his general report (page 118), summing up the situation in this regard, reports that the analysis of the national reports had revealed a reserved attitude to the trend that the OECD and individual tax authorities regard the prevention of double non-taxation as the purpose and objective of the interpretation of DTCs, and that the “existing DTCs based on the OECD model convention cannot be easily seen as serving the prevention of double non-taxation. There are hardly any indications in case law and scholarly legal writings suggesting that DTCs are pursuing this objective”.[4] In his presentation from the panel Professor Roy Rohatgi of India even went so far as suggesting that tax treaties based on the United Nations’ Model Tax Treaty, which had objectives mainly reflecting non-developed countries, were indeed intentionally construed to allow double non-taxation.
The discussion at the congress of whether the avoidance of double non-taxation is a purpose and object of tax treaties that should be taken into account in the interpretation process, seems to have been inconclusive.  Professor Daniel Gutmann, member of the panel from France, in commenting on the issue whether treaties should be interpreted extensively to achieve single taxation, or literally, to permit double non-taxation, determined that in his country the second mentioned literal approach is the rule, leaving no room for extensive interpretations of tax treaties. Professor Lang agreed with Professor Gutmann, noting also that because the object and purpose of tax treaties were to allocate taxing rights, the possibility of double non-taxation should be accepted. In his report from the congress in Svensk Skattetidning Mr. Jörgen Grönlund also reveals tacit support for a literalist approach to treaty interpretation.
Professor Hugh Ault, member of the panel from the US, seems to have been rather isolated, according to Svensk Skattetidning  in his opinion that one should adopt a more purposive and teleological approach of interpreting tax treaties in order to counteract double non-taxation situations. This, he maintained, was based on the general Vienna Convention rules on treaty interpretation. He also referred to the 1927 League of Nation report that every effort should be made to avoid double non-taxation and to ensure that all income be taxed once, and only once.
With regard to the conflict between literal and extensive (or purposive) interpretation it is worthwhile keeping in mind that the literal approach, supported by professor Gutmann, has the disadvantage not only of giving rise (more easily) to double non-taxation but also to double taxation. Also, one should avoid the term “literal” interpretation. Under article 31 of the Vienna Convention it is the “ordinary” meaning that should be explored, which, it is suggested, allows for a less rigid approach especially considering that this ordinary meaning should also take account of the context in which the term appears and of the purpose and object of the treaty.  Incidentally, it is regrettable that nowhere in the directives of the subject distributed to the national reporters nor in the general report is there a single reference to the Vienna Convention on the Law of Treaties.
Regarding Professor Lang’s statement that it is the purpose and object of a tax treaty to allocate taxing rights I respectfully disagree. The allocation of taxing rights under the treaties is not in itself the object of the treaty but only the means whereby, on a reciprocal basis, double taxation and indeed also double non-taxation is to be avoided.
Whatever the situation may be regarding the discussion of object and purpose, it seems clear that if a treaty includes a subject to tax or  a subject to remittance or any other similar provision clearly demonstrating a clear intention and purpose to prevent double non-taxation, it must be the obligation of the interpreter to serve these intentions and give such clauses the widest applicability possible. This, it seems, is acknowledged also in the general report (page 88). More hereinafter.
Professor Ault, at the congress, also presented a spectrum of distinguishable double non-taxation cases: those clearly intended, those not intended but that can be avoided by an appropriate tax treaty interpretation applying the Vienna Convention, and those not intended but where the ordinary meaning of the language of the treaty does not permit an appropriate interpretation. This differentiation of the matter brings into focus, which has already been touched upon above, the most important aspect of treaty interpretation namely the underlying intentions of the treaty making states, and in this case their intentions with regard to double non-taxation. This underlying intention of the treaty making states is also something  which the Supreme Administrative Court in Sweden (SAC) has repeatedly determined to be the goal for all treaty interpretation efforts.
Into the first category, i.e. cases where double non-taxation is clearly intended, would certainly fall all kinds of tax-sparing provisions. This is self evident and needs no further explanation.  As far as Sweden and the other Nordic countries are concerned an additional and very clear example of an intention to uphold double non-taxation  exists in the subject to tax provision in the Nordic multinational convention. Under this rule (article 26 paragraph 2) the state where the taxpayer is resident may tax whatever kind of income to which the primary right of taxation has been given to the source state but where this state does not avail itself of this right under its domestic tax laws. This, as it appeared, would, however, have the undesirable consequence that all kinds of Finnish state pensions and other Finnish family and social benefits paid under the Finnish pension laws, all of which are to be taxed in Finland under the treaty (article 18 paragraph 1) but which are tax free under Finnish domestic tax law, would be taxable under the aforementioned article if derived by someone who had moved to and become a resident of one of the other Nordic countries. As this of course would be quite unfair the following paragraph 3 of the article provides for an exception in these cases from the subject to tax provision in paragraph 2, thus intentionally creating a desired double non-tax situation.
In the second category where double non-taxation was not intended and seemingly was avoided by an appropriate purposive treaty interpretation, one may mention a case cited in the Swedish national report, RÅ (=Regeringsrättens Årsbok) R78 1:22. The problem in this case was how to determine the 183-day period of personal services under the employment income article conducted by a Swedish resident in South Africa, in a situation where the relevant “fiscal year”, during which the income was derived, as far as Sweden was concerned was the calendar year but where in contrast South Africa’s fiscal year ran from 1 July until 30 June and where Sweden as the state of residence of the employee used the exemption method for relieving double tax. The period of personal services performed by the Swedish tax payer in South Africa straddled the 1st of July in such a manner that if the Swedish calendar year were adopted the income would escape tax altogether whereas, if the South African fiscal year were chosen, Sweden would have the sole right of taxation.  Considering that the term “fiscal year” was not defined in the treaty and there was no other contextual reason to determine the understanding of the term according to either of domestic law meanings, it would not have been surprising if the Court would have adopted the Swedish meaning under the article of the treaty corresponding to article 3.2. of the OECD model convention. But it didn’t. It concluded that the meaning of the fiscal year should be determined under the law of the country where the services had been performed, i.e. South Africa, which thus gave Sweden the right to tax. This – it  could be suggested – was intended to prevent a double exemption that would have occurred if the Swedish domestic law meaning would have been adopted. This solution also lives up to the idea put forth in footnote 2 above, that when the interpretation of a term allows for two different meanings, one leading to a solution contrary to the object and purpose of the treaty and the other consonant therewith, the latter should be chosen. (If in this case the period of services performed in South Africa by the employee had instead straddled the first of January, the somewhat surprising result would have been that Sweden would have been denied the right to tax. Instead, however, South Africa would have been allowed to tax, and again double non-taxation would have been prevented.)
Into the third category of cases considered by Hugh Ault would fall any kind of situation where the ordinary meaning is clear but still leads to double non-taxation. The aforementioned capital gains loophole in the Sweden-Belgium treaty probably is a good example hereof.
It is uncontroversial that all measures should be adopted to prevent both double taxation and unintended double non-taxation. With regard to this latter phenomenon it also follows from the semantics of the title of tax treaties and, frankly, also common sense, that instruments designed to do away with double taxation only stretch that far and not to complete tax exemption. It is only logical that a taxpayer who is not subjected to double taxation requires no tax treaty protection.  Double non-taxation is in all but very few and carefully regulated cases an unintended effect of treaty application usually stemming from changes of domestic tax legislation that have occurred after the conclusion of the treaty. More  hereinafter.  Contracting States thus do not deliberately include provisions that give rise to double non-taxation. In such light it seems awkward to suggest that treaty interpretation should not pay respect to the intention of the parties that double non-taxation should be prevented. Certainly therefore the intention of the parties not to let the treaty allow for double non-taxation, should be a part of the interpretation process at least as an overriding ambition.
This, however, is not enough when it comes to interpreting the treaty. This process also requires, as emphasized by Professor Lang, that in its context, there be a reflection of these overriding purposes of the treaty in the text itself.  And often this is simply not the case. For instance, the distributive rules of tax treaties, which, on a reciprocal basis, give the primary right of taxation to one of the contracting states and compels the other state to give relief here-from, are of such a technical nature that they reveal very little about the object and purpose of these rules for interpretation purposes. For this reason these rules will frequently include detailed definitions of  terms and expressions that may give rise to interpretation problems. The aforementioned case regarding the Sweden-South Africa treaty reflects, however, a quite constructive purposive interpretation also of a distributive treaty rule.
With regard, however, to other types of treaty rules such as the residence article (number 4 of the OECD Model), the arm’s length rule (10), the avoidance of double taxation article (23 A-B), the non-discrimination article (24), and certainly such special rules as subject-to-tax or subject-to-remittance rules or limitation of benefits articles, it is much easier to discern the underlying purpose of the rule and the intentions of the Contracting States.  In these cases a more purposive, extensive and teleological approach to the interpretation process is appropriate.
A quite specific reason why tax treaties, in contrast to domestic tax laws, should be allowed to be interpreted more purposively or extensively is because they are by nature quite different from domestic tax laws. Whereas such domestic tax laws, alas, may cover hundreds, not to say thousands of pages, a tax treaty   will only be about 30 to 40 pages long. They must also adapt to future changes of the domestic tax laws of the contracting states. Treaty interpretation is not a process that is carved in marble once and for all but an art that must conform to the realities of ever changing realities. All this calls for a more flexible, purposive and imaginative approach where in good faith one must constantly probe into what was the underlying intention of the parties.  The European Court of Justice is a strong adherent of such interpretation principles (see Maria Hilling’s recent dissertation “Free Movement and Tax Treaties in the Internal Market”), something, for which they in general have won great respect.
The Swedish national report concludes that the  object and purpose of tax treaties, referred to in their titles, has not been decisive for the SAC, as the result of its interpretation has led to both double taxation and double non-taxation. This statement (page 673), as far as double non-taxation is concerned, is based solely on the outcome of an advance ruling, RÅ 1996 ref. 84.  The problem in this case, however, was not focused on a double non-taxation issue but to establish, for the purposes of applying the Swedish domestic CFC-legislation, if a Luxembourg company was “liable to tax” and thus if it was a resident for treaty purposes in Luxembourg, which was, at the time, a condition for non-application of the Swedish CFC-legislation. The Court found that the Luxembourg company, which was subject to a very insignificant tax (which, in addition, may not have been covered by the treaty under article 2), was nevertheless “liable to tax” under article 4 of the treaty and thus a resident of Luxembourg, rendering the Swedish CFC-legislation non-applicable. The question whether the Luxembourg company could benefit from any kind of double non-taxation was, however, never raised in this case. Considering that the company in question only earned passive income there appears to be no situation in the treaty where Sweden as source state had fully given up any taxing rights. Or, in other words, the Luxembourg company  would never benefit from any double non-taxation benefits just because it was deemed a resident of Luxembourg by the SAC. It is therefore not possible to draw any certain consequences of the Court’s attitude to the interpretation of double non-taxation matters in this case.
(Quite generally the rationale for the exploration of the “liable to tax”-test in the context of double non-taxation seems a bit obscure. If the answer to this question is negative, i.e. the person in question is not a resident of either state and thus not within the personal scope of the treaty, a possible double non-taxation of his income will not be the result of an application of the treaty but due to the domestic laws of the two countries. If on the other hand the answer is positive, i.e. the person is liable to tax in one of the contracting states and thus within the scope of the treaty, he can of course in principle, just like any other person, be entitled to double non-taxation under the treaty. But this is not because he is liable to tax in one of the countries but because of the interaction of a specific treaty clause and domestic law as the case may be.)
The Luxembourg case, RÅ 1996 ref. 84, will also be discussed below with regard to “subject to tax”– considerations.
The Swedish national reporter does, however, have support for her conclusion  (page 674) that the object and purpose to counteract double non-taxation has not been decisive for the SAC. And this, which is most important to observe, follows from their interpretation  - in two cases - of specific anti double non-taxation clauses! Or in other words, the interpretation by the SAC of treaty clauses specifically designed to prevent double non- taxation  has nevertheless led to such double non-taxation. The first of these cases dealt with the subject to remittance rule in the former Sweden-UK treaty (RÅ 1987 ref. 162). [5] The second case addresses the subject to tax clause in the Sweden-Peru treaty (RÅ 2004 not 59 [6]).  The Court, despite the fact that they in general terms (and with great eloquence) in the first mentioned case, and again in RÅ 1996 ref. 84, have solemnly proclaimed their faith in the Vienna Convention rules, that one shall above all explore and respect the intention of the contracting parties, have nevertheless in both rulings interpreted these specific anti non-taxation clauses very narrowly, giving rise in both cases to double non-taxation. The SAC have thus demonstrated that the prevention of non-taxation is very low indeed on their list of priorities and that they in fact seem fettered to their traditional textual approach to interpretation of domestic law. 
As an example where a ruling by the SAC gave rise to and thus accepted double taxation as a result of their treaty interpretation (page 659), the Swedish national report refers to RÅ 2001 ref. 46. This conclusion is not only correct, it is also a major  understatement! But, more importantly, the case represents an embarrassing blunder by the Court with regard to their understanding of the relevant article of the Sweden-US treaty and of the tax principles in general of taxation of foreign partnerships. Any international tax analyst would take a 100-1 bet that a renewed hearing by a full seated Court (in pleno) would reverse the outcome of this decision. It raised an outcry in the Swedish international tax community and the exhaustive analysis of the case in Jesper Barenfeld’s recent dissertation “Taxation of Cross-Border Partnerships, Double-Tax Relief in Hybrid and Reverse Hybrid Situation” is not panegyrical. The ramifications of the case were indeed so serious that the legislator had to step in - but it took more than four years –  and change our domestic credit of foreign tax law.


How does double non-taxation arise?

This question is valid because it may give an answer for determining whether a double non-taxation situation is intended or not. Another reason for exploring the underlying reasons for the double non-taxation situation, a reason which is also mentioned by the national reporter of the United States (Diane M. Ring), is that the “identification of the cause of the non-taxation is essential to any effort to evaluate possible responses”.
Maybe, already at the time of the conclusion of the treaty, it was possible to foresee the double non-taxation, and still nothing was done about it. Is it then not fair to say that such double non-taxation was intended or at least acceptable? In such a case, where thus the negotiators were aware of the double non-taxation, it is more difficult to maintain the position that double non-taxation was unintended.  This appears to have been the case summarized by Professor Rohatgi at the congress regarding the double non-taxation of capital gains that followed from a Mauritius’ holding company divestment in India. The situations where double non-taxation exists already at the conclusion of the treaty, and is left unopposed, are probably quite few. One such case, however, as far as Sweden is concerned, was the abovementioned treaty with Belgium where Sweden, as situs state, gave away its right to tax capital gains on sales of Swedish real estate, where, at the same time, Belgium did not tax foreign gains.
Another typical situation where double non-taxation can be foreseen is when a treaty is made with a country that applies a territorial tax system. If, in this case, the residence state adopts the exemption method for relief of double tax, third country income will always escape tax altogether. In its treaties with such countries Sweden will therefore always include a subject to tax rule. Also where double tax is avoided by the credit method a double non-taxation situation may nevertheless occur with regard to third state income. For instance, in its treaty with South Africa, which at the time of its conclusion (1995) had a territorial tax system, Sweden included a quite sophisticated subject to tax rule (article 26 paragraph 2) regarding third state income. In this treaty Sweden relieves double tax by way of credit of South African tax but may nevertheless impose taxation of third state income derived by a resident of South Africa if he/she is also resident in Sweden for domestic tax purposes and the income has not been taxed in South Africa. The example given in the government bill presented to parliament explaining this rule is a tax- payer who is resident in South Africa under the treaty but resident also in Sweden under domestic Swedish tax law and who sells immovable property located in a third state. The gain from the sale is tax free in South Africa under its territorial system but may be taxed in Sweden under the subject to tax rule.[7]
No, the most common reason for double non-taxation, and which has already been mentioned, is that one of the Contracting States changes its domestic law after the conclusion of the treaty.  Another, and equally unforeseen situation, could be that some kind of tax planning device has been construed which was not possible at the time of the conclusion of the treaty or that some kind of financing or other economic invention for deriving income has been developed. In these cases where the reason for the double non-taxation situation emerged after the treaty was made, one can truly establish that this was unintended.
If thus in conclusion the double non-taxation situation existed and could have been foreseen already at the time when the treaty entered into force, then one could suggest that this result was accepted or intended. Where, on the other hand, the double non-taxation situation emerges from a development that occurred after the conclusion of the treaty, usually due to a change of domestic law in either of the countries, then the double non-taxation situation is unintended allowing for a more purposive interpretation approach. If this is not possible the only way of preventing the double non-tax situation is to renegotiate the treaty.
Double non-taxation in domestic law.
It could also have been useful in the IFA reports to have  explored more deeply to what extent double non-taxation is a problem in domestic legislation and to draw conclusions from the efforts made in the reporting states to curb such practices. The Swedish experience in this regard could have been quite instructive for the IFA community.
Sweden, in its role as state of residence of the taxpayer,  is a country that applies in its domestic tax legislation the principle of world wide tax liability  with credit for foreign taxes, thus allowing  very little room for double non-taxation. There is, however, one situation, which is also very common, where Sweden exempts foreign income from tax, namely employment income derived by Swedish resident employees expatriated for more than six months, the so called six-months rule. This situation, for a number of years, gave rise to frequent double non-taxation opportunities. The main thrust of the rule is now, after several rounds of legislation that the taxpayer must prove that the income has been reported and assessed for tax purposes in the other state before exemption is allowed in Sweden. This has had the positive effect that litigation in this field has all but ceased. This rule will be discussed also below under the subtitle dealing with suggestions for counteracting double non-taxation.
A further experience of a domestic double non-taxation situation in Sweden worth mentioning, albeit an economic double non-taxation situation of corporate profit, occurred in 1994 when a Swedish right-wing government, (facing an expected loss at the polls of the upcoming general election), suddenly introduced an exemption from taxation on dividends received by all Swedish individual residents from Swedish companies thus doing away with our longstanding principle of (classic) double taxation of company profits. This led to a big problem with our tax treaties allowing for an unintended extension of the exemption also to dividends received from foreign companies by interposing a Swedish holding company of the foreign shares. This loophole, however, was immediately recognized and led to the introduction of an extremely complicated piece of legislation (vidareutdelningsskattelagen) whereby the proportion of dividends paid to the Swedish shareholder by the Swedish holding company, that was attributable to dividends derived from the foreign companies, would nevertheless be taxed when distributed to the Swedish shareholder of the holding company. When illustrated as a diagram the tax law resembled something like the London underground system! Much to the relief of Swedish tax practitioners (or to the chagrin of the real experts on this law who could expect to collect substantial consulting fees) the law was changed back after only one year when the re-elected social democratic government re-introduced double taxation of dividend income.
The two most notable types of income where Sweden as the state of source of the income grants an exemption from (withholding) tax, regardless of whether the foreign person deriving the income is exempted from tax also in his home state or not, are interest income[8] and capital gains on movable property including of course stock holdings. This, however, is a result of general tax policy aimed at promoting foreign investments in Sweden. (Sweden can thus not enforce the 10 percent withholding tax on interest normally granted to the source state under its tax treaties.)


What types of income are mainly prone to double non-taxation exploitation?

This question is also valid as it may give an answer to what measures should be adopted in tax treaties to prevent double non-taxation.
The answer as far as Sweden is concerned is quite simple: capital gains.  In almost all the aforementioned cases where double non-taxation has occurred in Sweden on a large scale this has involved capital gains.  The reason why double non-taxation of capital gains is especially serious is of course that this type of income many times will involve very great sums of money and that the income is generated immediately. This makes it worth while for the seller, if this is necessary, to move even temporarily to the pertinent tax treaty state just to collect the double non-tax benefit. One method for counteracting such tax planning schemes is to introduce the type of carry-back tax liability introduced in the UK where a taxpayer who moves abroad and then moves back to the UK within five years will be liable to tax on his return for whatever gains that he made during his stay abroad. But certainly, these situations should be especially observed in subject to tax clauses in treaties.
Other forms of double non-taxation resulting from tax treaty application.
The reports to the congress have mainly addressed the double non-taxation situation that occurs where the Contracting State that has been afforded the sole taxing right of the income cannot in effect make use thereof under its domestic law. There are, however, other forms of double non-taxation emerging from the interaction between treaties and domestic law, which would have merited further examination at an IFA congress. In Sweden, for instance, a favourite tax planning ploy for Swedes that permanently move abroad is to maintain their Swedish residence under domestic law, which is quite simple to do because residence continues automatically for five years unless the taxpayer proves otherwise. The double residence situation which thus normally occurs under the tie-breaker of the treaty will have the effect that the emigrant lender, under the treaty, is relieved from Swedish tax on his interest income but at the same time, under domestic Swedish law, preserves his entitlement in Sweden to a deduction for his worldwide interest expense, irrespective of the fact that the same interest may be deductible in the other Contracting State too.  Usually the double non-taxation effect in these cases thus appears as a double deduction for interest expense.  Thirty percent of the net interest loss can then be transformed into a credit of tax, which can be offset against other taxes, most conveniently against the abominated Swedish real estate tax on the summer house which the expatriate often will retain in Sweden. This applies also, which is a common situation too, where the interest expense is attributable to loans taken for the acquisition of the taxpayers new home abroad! (In effect this tax-planning device means that even the tax base of Swedish immovable property can be transferred abroad!)
Subject to tax (STT) clauses.
From the summary of the break-out session in Vienna it appears that there was a lively debate and many suggestions regarding subject to tax clauses of various forms. The national and general reports have of course also covered these matters illustrating inter alia the difficulties that follow from different understandings of such terms as “liable to tax”, “subject to tax”, “taxed” and “effectively taxed”.
In Sweden, in addition to the contribution made by the Swedish national reporter, there has been a significant scholarly debate over the aforementioned Supreme Court tax case RÅ 1994 ref. 84 regarding the understanding of the term “ liable to tax” in its context in the residence article 4 of the former Luxembourg treaty. (I, too, have participated therein and it has also attracted the interest of Mr. Jörgen Grönlund in his report from the Vienna congress.) It should be mentioned, however, that this case is of limited interest with regard to the issue of double non-taxation because Sweden does not use the “liable to tax” test as a condition in its treaties for anti double non-taxation purposes but always uses the term “subject to tax” in these situations. This means that the fact that a person in the other Contracting State is “liable to tax” will not save it from the “subject to tax” clauses, which are used in Swedish treaties to prevent double non-taxation.
As mentioned above Sweden has had positive experience with regard to “subject to tax” provisions in its domestic six- month rule. As already mentioned the condition for a tax exemption in Sweden under this rule is that the foreign earned employment income has been reported and assessed to tax in the other state.  There is no minimum requirement, however, regarding the level of the tax rate imposed in that state nor that any tax has been effectively levied, which may happen if e.g. there are minimum income thresholds, special allowances or if a loss carry-forward eliminates taxation. The taxpayer must only demonstrate in Sweden that he has reported the income to the tax authorities and thus been subjected to assessment, which in most cases is quite simple by just providing a copy of his foreign tax return. The advantages of this regime are obvious, a) in that it makes it unnecessary for the court to investigate the domestic tax laws of the other country, b) there is no need, as the case may be, for any certificates of residence and c) it automatically catches all the tax cheaters that may have evaded taxation in the other country by simply not reporting their income. A recommendation by IFA to consider such an approach also in tax treaties could have been valuable.
A further problem regarding subject to tax and subject to remittance rules, which could have merited an investigation by IFA, is how to determine whether the income is in fact subject to tax or subject to remittance in the other state and who has the burden of proof for this. In Sweden income tax treaties (in contrast to inheritance and gift tax treaties) are implemented ex officio but it is up to the taxpayer to prove that the conditions for treaty relief are fulfilled. The taxpayer should in consequence always be obliged, as the case may be, to demonstrate that the income is subject to tax under the laws of the other state or that the income has been remitted before relief is provided. These problems, it appears, are treated rather lackadaisically in Sweden. Anyway, there are no regulations or guidelines for employers, financial institutions, insurance companies or other paying agents of how to handle these problems for withholding of preliminary tax purposes and the tax authorities in a world of deregulated and “cyber-spaced”  financial markets have very limited resources for monitoring these problems.
Further suggestions for counteracting double non-taxation.
Several Cahiers reporters, in a rather defeatist mood,  conclude that double non-taxation is something that must be “accepted”. But surely IFA must demonstrate a more activist and determined stance! And indeed it seems that the discussion at the break-out session focusing on subject to tax clauses for counteracting double non-taxation was quite intense and productive.
As mentioned above double non-tax situations usually occur due to changes of domestic laws that take place after the conclusion of the treaty.  In order to put a stop to these situations, if they are considered unwarranted, it is important that contracting states make more use of the provision in all treaties, usually article 2.4, to inform each other of changes in their domestic laws that affect their treaties, so that measures can be taken as soon as possible to plug the loophole.[9] One could also consider introducing a shorter period for terminating tax treaties than the six months stipulated in article 30 of the OECD Model and/or to devise partial termination of treaties. This could be especially helpful in cases of double non-taxation of capital gains. Retroactive legislation is banned in Sweden but it is possible for the government, in exceptional cases of urgency, to officially inform parliament of a planned change of legislation and at the same time  to decree that the change will be made effective as of the date that the information was submitted. Some kind of similar device could be considered for tax treaties too.
The kind of double tax relief that occurs in Sweden by granting deductions for interest expense under national law, while at the same time the right to tax interest income is barred by the treaty, is of course easily cured by just changing the first mentioned domestic law stipulating that a deduction is denied for interest expense where taxation of interest income is waived under a treaty. This, as a matter of fact, was done in the past, but only with regard to interest attributable to business income. Thus, Swedish banks that derive tax free interest income under certain (but now very few) exemption treaties may no longer deduct interest expense or other costs, attributable to the (tax exempt) income.  Also, which occurs in Sweden in cases of double residence under tax treaties when someone has in fact moved abroad, it seems illogical and unwarranted to continue allowing general deductions and deductions e.g. for premiums on Swedish pension schemes, which are granted only to residents of Sweden, all of which may give rise to double exemption benefits. An alternative is to decree that interest expense and other deductions allowed to residents of Sweden apply only if such deductions are not taken in any other jurisdiction.
The main problem with STT clauses in tax treaties, which is the standard measure to deal with double non-taxation, is that the double non-tax opportunity, as mentioned above, frequently appears (a long time) after the conclusion of the treaty and that there will be no STT clause to deal therewith. Thus one will have to renegotiate the treaty, which is a long and cumbersome affair, leaving the double non-tax loophole open for a long time. In some cases where there is an STT clause it may not cover the situation at hand or give rise to interpretation problems. In order to cure these drawbacks, it could be worthwhile considering the introduction of a domestic STT clause. This is of course formally an override provision. But it is a harmless one! Most importantly, and keeping in mind that treaties are agreements between states, such a rule does not affect the taxing rights of the other state as this has no tax claims under domestic law. Usually, also, the non-tax situation has emerged due to a change of legislation in this other state thus changing the underlying basis for the original negotiation of the treaty. This is something that provides a weak moral base for complaint about the override provision of the other treaty partner. But most importantly,  such an all-encompassing domestic STT rule serves the legitimate purpose of  coming to terms with  unsatisfactory, unintended, unwarranted, unfair and unforeseen effects of the treaty, all of which, in the final analysis, is in harmony with the general principles of the treaty and the basic intentions of the contracting parties. Moreover, such a domestic STT-rule does not encroach upon any legitimate or fundamental rights of the taxpayers and upholds the universally agreed norm that all income should be taxed once and only once.
When Sweden is the state of residence of the taxpayer under its treaties there are very few opportunities for double non-taxation. The reason for this, which has already been mentioned above, is that Sweden in almost all its treaties adopts the credit of foreign tax method for relieving double taxation, which means that relief is only given for taxes actually paid in the other country. Also, in a few of its treaties, Sweden operates a system of reverse credit, which, where Sweden consequently is the source state, will take care of situations where double non-taxation may arise. In this sense the reverse credit has the same effect as an STT rule which does not affect the tax situation in or the tax policy of the other state. In the Nordic multilateral tax treaty Sweden, as state of residence of the taxpayer, uses the exemption method for double tax relief regarding employment income, a measure which of course affects a very substantial number of taxpayers. When this was introduced it gave rise to such widespread and aggressive tax planning that a switch-over clause was subsequently negotiated whereby Sweden, upon a simple diplomatic notification procedure, can switch back to the credit method. (According to the congress report such switch-over rules were indeed discussed at the congress proceedings.)
Concluding remarks.
Sweden’s tax treaties today, for reasons which, however, are too comprehensive to explain in this context, have, to a very large extent, played out their role as instruments for avoiding double taxation at least as far as Swedish resident taxpayers are concerned. For residents of our treaty partners our treaties have also lost interest as most favours granted therein, especially in terms of reduced Swedish source taxes are now granted to everybody under domestic Swedish tax laws, i.e. also to residents of states that do not have a treaty with Sweden. Indeed, a Swedish treaty negotiator once exclaimed that we must retain our tax claims so that there would be some taxes left to surrender in our treaty negotiations. Otherwise there would be no interest for other states to make tax treaties with us!
Sweden, like most other treaty making states, is of course constitutionally obliged to abide by its tax treaties, but only to the extent that such application restricts tax liability under Swedish domestic tax law. For tax-payers burdened by the highest income taxes in the world this naturally has a great attraction! Every effort is thus made to explore the possibilities of loopholes or other benefits in our treaties, including double non-taxation.
Every serious analyst will agree with Hugh Ault and the founders of the League of Nations almost a hundred years ago that every effort should be made not only to prevent double taxation, but also to ensure that income does not go untaxed altogether. International agreements carefully negotiated by governments for the purpose of avoiding double taxation that give rise to situations where no tax is charged at all and thus emerge as instruments for double non-taxation, in the eyes of the ordinary tax payer, serves to make a mockery of the tax system in general and of tax treaties in particular. In certain treaty interpretation situations, as pointed out by the general reporter in Vienna, we may well have to “accept” double non-taxation, but that is certainly not the norm as far as tax policies are concerned.
The International Fiscal Association has always been a champion of the crusade against double taxation. It is important for its integrity that it demonstrates equal determination in the quest to counteract also double non-taxation.
(IFA member 2709)

[1] Even more surprisingly, the Swedish government bill regarding the Belgium treaty did not even inform the Swedish Parliament about this quite spectacular deviation from ordinary treaty policy. It is difficult to determine the reason for this surrender of tax by the Swedish negotiators. Maybe one reason could be that, at the time of the conclusion of the treaty, Sweden enforced quite stringent restrictions on foreign ownership of Swedish real property and/or that capital gains tax on such property  was due only during a ten year ownership period.
[2] Correspondence available upon request, (Swedish only).
[3] Correspondence (in English) available upon request.
[4] In Austria itself it appears, however, that there is no total consensus in this matter. Thus, the national reporter for Austria (Heinz Jirousek) declares (page 196): “Within the framework of an interpretation method that refers to the object and purpose of an international agreement, and on the basis of court decisions at the highest level and with due consideration of the principles developed by the OECD, the Austrian tax administration proceeds from the view that both double taxation and double non-taxation as a consequence of treaty application should be incompatible with the purpose and object of a double taxation treaty.”  And also (page 173): “Aus der Sicht der österreichischen Finanzvervaltung gilt als grundsätzliches Auslegungsprinzip, dass immer dann, wenn der Wortlaut einer Abkommensbestimmung zwei Auslegungsvarianten zulässt, wobei die eine den Eintritt der Doppelbesteuerung oder Dobbelnichtbesteuerung nach sich zieht, während die andere dies vermeidet, dieser anderen Auslegungsvariante der Vorzug einzuräumen und die erstgenannte als unrichtig zu verwerfen ist.”
[5] An English analysis of this case has been given by this author in British Tax Review 1990 no 9 (Interpretation of Tax Treaties – a case study).
[6] This judgement was published in 2004, after the Swedish National Report had been submitted. I therefore sent the case to the discussion panel for their consideration. It will be the specific topic of a forthcoming article by this author.
[7] Since this treaty was concluded in 1995 South Africa has switched (in 2001) to a world wide tax liability system making this subject to tax rule obsolete.
[8] Not only does Sweden not impose any tax on outgoing interest payments, it also allows, in the corporate sector, for a full deduction of interest expense. And there are no thin capitalization restrictions. All of this makes Sweden an ideal jurisdiction for establishing international holding companies. In the past, the Advance Rulings Board in Sweden, issued numerous and repetetive rulings, which were never appealed by the tax authorities, confirming the deductions for interest expense also in thinly capitalized Swedish subsidiaries of foreign corporations, but only on the condition that the deducted amount was subjected to taxation by the foreign recipient. This condition thus prevented a double tax relief of the interest amount.
[9] The Swedish competent authority once claimed with regard to our Japanese treaty partner that they had fulfilled this obligation by sending to Japan the complete Swedish tax code - in Swedish!