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 privat hjälp av en skattekonsult.

 

Interpretation of tax treaties authenticated in two or more languages – a case study. [1]

 

 

 

 

In 2004 the Supreme Administrative Court of Sweden (SAC), delivered an advance ruling, RÅ 2004 not 59, regarding the interpretation of the Sweden – Peru tax treaty that was signed 17 September 1966.

 

The case focuses on the problems of  treaty interpretation caused by a divergence of language in the two texts of the treaty,  but it highlights also a number of additional interpretation issues.

 

Swedish bilateral tax treaties are generally drawn up in the languages of the two  contracting states both languages being authentic and thus equally authoritative  for interpretation purposes. Sometimes, for instance in the treaty with China, there is also a third authentic text  (usually in English) which, in case of doubt, shall prevail. A few treaties, for instance those with the US and France, are only in one authentic language, English and French respectively. The multinational Nordic treaty is authenticated in seven languages (including two Swedish texts, one for Sweden and one for Finland.)

 

In contrast to most other advance rulings the files of this ruling with all its documents have not been  classified, which  has allowed for a reconstruction of the exact proceedings and administrative management of the case, parts of which are quite surprising!

 

The ruling  has generated a lively tax planning debate in Sweden and given rise to such wide-spread abuse of the treaty that the Government have terminated it!

 

 

------------------------------------------------------------------------------------------------------------------------------

 

Table of Contents:                                                                                                                    Page

 

FACTS OF THE CASE…………………………………………………………….……...….2
APPLICABLE DOMESTIC TAX LAWS………………………………………… ……..…2

APPLICABLE PROVISIONS OF THE SWEDEN-PERU TAX TREATY………     ……2

THE RULING OF THE ADVANCE RULINGS BOARD (ARB)…………………..… ..…3
THE RULING OF THE SUPREME ADMINISTRATIVE COURT (SAC).…………. .…4
ANALYSIS OF THE DECISIONS OF THE ARB AND THE SAC…………….…….. ….4
        The ARB……………………………………………………………… ……………… …4
        The SAC……………………………………………………………… ……….……… …5
           Diverging treaty texts and article II §2 (the lex fori rule)……….. ………..……… ..9
           The meaning of  “income” and “ from a source” ………………… ………  ……….9

            The meaning of “réditos de fuente” in the Spanish version of the treaty..….……..12

           Supplementary means of interpretation according to article 32 of the Vienna

          Convention……………………………………………………………………   ………13

          The object and purpose of the treaty…………………………………………   ……..15

           Double non-taxation……………………………………………………………  …….18

           Structure of article XVII of the Sweden-Peru treaty………………………… …….18

           Reconciliation of the texts………………………………………………………  ……20
TREATIES WITH LESS DEVELOPED COUNTRIES AND TAX SPARING…..  .....…22
ARTICLE  3.2. (of the OECD model) AND TREATY PROVISIONS APPLICABLE
BY ONLY ONE OF THE CONTRACTING STATES……………………………………. 22
IN DUBIO, PRO TAX PAYER OR PRO FISCUS?…. …………………………………... 23
CASE RÅ 1987 ref. 162…………………………………………………………………  ..….24
CONCLUDING REMARKS……………………………………………………………   ….26

 

 

 

 

FACTS OF THE CASE:

 

The facts of the case were quite simple: Two Swedish companies jointly owned all the shares of two Peruvian companies. The Swedish companies, intended to sell all their shares and the question that arose was simply; where, under the treaty, shall the capital gain be taxed, in Sweden or in Peru?

 

 

APPLICABLE DOMESTIC TAX LAWS

 

At the time of the sale of the shares the gain, under Swedish law, constituted taxable income.[2] Under Peruvian tax law, as explained by the applicant Swedish companies, the matter of capital gains taxation was not  quite clear but information provided by Peruvian tax counsel had indicated that there would be no capital gains or other tax consequences in Peru. The courts therefore stipulated that for the purposes of the ruling it was assumed that the gain was not subject to taxation in Peru.[3]

 

 

APPLICABLE PROVISIONS OF THE SWEDEN-PERU TAX TREATY:

 

Article X

 

(Swedish text):

Vinst, som härflyter av försäljning, byte eller överlåtelse av förmögenhetstillgång (vare sig denna utgöres av lös eller fast egendom), får beskattas allenast i den avtalsslutande stat, i vilken tillgången är belägen vid tidpunkten för försäljningen, bytet eller överlåtelsen i fråga.

 

(Spanish text):
Las ganancias de capital obtenidas de la venta, transferencia o permuta de bienes de capital, sean muebles o inmuebles, solo podrán gravarse en el territorio del Gobierna Contratante en que tales bienes, hayan situados en la época de la venta, transferencia o permuta.

(English non-authentic text  by the International Bureau of Fiscal Documentation:)
Capital gains derived from the sale, transfer or exchange of capital assets, whether movable or immovable, shall be subject to tax only in the Contracting State in which the assets were situated at the time of such sale, transfer or exchange.

Article XVII, §2

 

(Swedish text):

Där icke bestämmelserna i artikel VIII annat föranleda skall inkomst från inkomstkällor i Peru, vilken inkomst enligt peruansk lag och i överensstämmelse med detta avtal är underkastad beskattning i Peru vare sig direkt eller genom skatteavdrag, vara undantagen från svensk skatt.

 

(Spanish text:)
Con sujeción a las disposiciones del articulo VIII, el rédito de fuente en el Peru que de acuerdo con las leyes del Perú y de conformidad con este Convenio está sujeto a impuesto en el Perú, ya sea directamente o por deducción, estara eximido del impuesto Sueco.

(English non-authentic text by the International Bureau of Fiscal Documentation:)
Subject to the provisions of article VIII, income from sources in Peru which under the laws of Peru and in accordance with this Agreement is subject to tax in Peru either directly or by deduction shall be exempt from Swedish tax.

 

THE RULING OF THE ADVANCE RULINGS BOARD, ARB, (André, (chairman), Hallberg, Eriksson, Ståhl):

 

The ruling by the ARB was based only on the Swedish text of the treaty. The Spanish text had not been consulted and not included in the case dossier.

 

The relevant parts of the ARB ruling translated into English are as follows:

 

“According to article X “capital gains derived from the sale, transfer or exchange of capital assets, whether movable or immovable, shall be subject to tax only in the Contracting State in which the asset is situated (emphasis added) at the time of such sale, transfer or exchange.

 According to the wording of article X there is nothing that indicates that gains from the sale of securities should be excluded from its application, albeit that the expression that the asset be situated (in one of the states) can be difficult to apply with regard to the shares of a company. The ARB has found no support for the idea that it was intended by the contracting parties that a sale of  shares should not be encompassed by article X.

A consequence of the opinion that article X thus is applicable to a  sale of shares is that they can be considered situated in the state in which the seller is not resident. This requires that other elements of attachment (anknytningsmoment) can justify such a conclusion. This could be the place of incorporation of the company or the place where the activities of the company or the circumstances giving rise to the appreciation of the company have occurred. The structure of the treaty (see also below regarding the interpretation by the ARB of article XVII) and the considerations underlying the allocation of taxing rights of the various types of income weigh heavily in favour of considering the elements of attachment to be allocated to the source state.

 

Considering that the companies are registered in Peru and that their business activities have taken place solely in Peru, the ARB is convinced that the shares of the Peruvian companies for the purposes of article X are situated in Peru. Therefore the gains according to this article shall be subject to tax only in Peru”.

 

With regard to article XVII § 2 the ARB concluded:

 

“According to the opinion of the ARB the wording of the article indicates that it is a subject to tax-rule, i.e. that the right to tax reverts to Sweden if the conditions of the article are not met. To support this view is the fact that article XVII § 2 would be rendered meaningless if it were not given this meaning, as already the distributive rules of which article X is one, gives the primary right of tax only to the source state.

 

Therefore, and assuming that the gains from the sale of the Peruvian shares will not be taxed in Peru under Peruvian domestic laws, Sweden, according to article XVII, may subject them to tax.”

 

Two members of the ARB (Wingren and Brydolf) also found in accordance with the majority, that the shares of the Peruvian companies were “situated” in Peru according to article X. They concluded, however, that the income from the sale of the shares did not arise “from a source” in Peru in the meaning of article XVII § 2 which followed from the principal structure of the treaty, according to which, as a general principle, Sweden, for the relief of double taxation, had adopted  the exemption method. Reference was here made to Government Bill 1967:26 s.25 (introducing the treaty to Parliament).

 

One member of the ARB (Virin) came to the conclusion that the Peruvian shares under article X were situated in Sweden giving Sweden the primary right to tax the gains. This had the consequence that there was no need to contemplate the application of article XVII §2.

 

 

THE RULING OF THE SUPREME ADMINISTRATIVE COURT, SAC, (Susanne Billum, Gustaf Sandström, Stefan Ersson[4], Eskil Nord, Carina Stävberg – unanimous):

 

Upon appeal and when the case was presented to the judges of the SAC the Spanish text of the treaty  had also been included in the file. Still, however, no detection had been made of any differences between the Swedish and Spanish language versions. This was discovered only at a very late stage, see below, and led to a dramatic reversal of the outcome in the ARB.

 

The SAC agreed with the ARB in its finding and its reasoning that, for the purposes of article X, the shares of the Peruvian companies were “situated” in Peru thus assigning the primary right of taxation of the gains to Peru. With regard to article XVII 2 §, however, the Court declared:

 

“The treaty has been drawn up in the Swedish and Spanish languages. The effect hereof – when, as in this case, nothing to the contrary has been indicated – is that both language versions are equally authentic. When applying the treaty in Sweden the Swedish text shall primarily be used (see article II § 2 of the treaty). In cases of uncertainty, however, the Spanish text too shall be considered. According to article X, capital gains derived from the sale, exchange or transfer of a capital asset (las ganancias de capital obtenidas de la venta, transferencia o permuta de bienes de capital) may be taxed only in the contracting state in which the asset is situated at the time of the sale, exchange or transfer in question. According to article XVII §2, income from sources in Peru, subject to the provision of article VIII, which under Peruvian law and in accordance with this Convention is subject to tax in Peru either directly or by deduction, shall be exempted from taxation in Sweden. According to its wording this latter article may be considered a general so called subject to tax-rule, i.e. a rule under which Sweden has the right to tax if Peru does not exercise its taxing rights according to the treaty.

 

The question whether or not the term ”income from sources” is intended to comprise capital gains cannot be determined solely from the Swedish text. In comparison to the Spanish text a difference emerges in the expressions. The Swedish text “inkomst från inkomstkällor” in article XVII § 2 is thus in the Spanish text represented by “el rédito de fuente” which verbatim means “avkastning från källa”. As thus the Swedish text does not answer the question whether the provision is intended to cover capital gains, this, from an examination of the Spanish text, appears as very unlikely. From the enumeration of the Peruvian taxes covered by the treaty (article I §1,2) it also appears that Peru, at the conclusion of the treaty, levied taxes partly on income (impuestos sobre la renta), partly on capital gains derived from the alienation of real property (el impuesto a las ganancias de capital provenientes de la transferencia de immuebles). As the structure of the Spanish text virtually excludes article XVII §2 from covering capital gains it is thus established that article X of the treaty is intended exhaustively to determine which state has the right to tax a capital gain where a person, resident in one of the contracting states, sells an asset in either of these. This means that the right to tax the capital gains in question is assigned to Peru only and that it is immaterial whether the gains have been effectively taxed in Peru or not”.

 

 

ANALYSIS OF THE DECISIONS OF THE ARB AND THE SAC.

 

As gleaned from the above reports of the decisions by the ARB and the SAC, there is, save for one vote in the ARB, full unanimity with regard to the interpretation of article X determining that the shares of the Peruvian companies were “situated” in Peru. Peru thus has the primary and sole right to tax the gain under article X.

 

The following analysis is therefore only concerned with the second question that arose, i.e. the extent to which the subject to tax rule in article XVII § 2 was applicable, giving Sweden the right to tax.

 

 

The ARB:

 

As already mentioned the decision by the ARB was based solely on the Swedish text of the treaty and the problems stemming from the wording of the Spanish text never arose as this text had never been consulted. Thus, without even bothering to motivate its conclusion, the ARB considered the gains to be income from a source in Peru and because it had not been subject to tax in Peru article XVII § 2 gave Sweden the right to tax the gain.

 

The decision reached  by the dissenting members of the ARB, Wingren and Brydolf, regarding article XVII § 2 is a little bit more problematic. As mentioned, they came to the same conclusion as the majority of the ARB in determining that, according to article X, the Peruvian shares were situated in Peru, because the structure of the treaty and the deliberations underlying the allocation of taxing rights of the various types of income etc. weighed heavily in favour of considering the elements of attachment to be allocated to Peru.

 

But with regard to article XVII § 2  the income derived from the alienation of the same shares was not considered to have its source in that country. This gives rise to the question: Where then does the income from the sale have its source? Apart from Peru there are only two alternatives, either that the gains are derived from a source in Sweden or from a source in a third country. Of these alternatives the latter must no doubt be immediately ruled out. Only Sweden remains.  Thus the income derived from the sale of shares which for capital gains tax purposes in the treaty were “situated” in Peru  had its source in Sweden!

 

It should be noted, however, that Wingren and Brydolf, just like the rest of the ARB, also found that the gains constituted an item of income.

 

See below for a further discussion regarding the source of an income.

 

 

The SAC:

 

At the very first inspection of the dossier of the case it was noted that the memorandum (föredragningspromemoria) provided by the Court Secretary (Magdalena Wetterfors) to the judges of the SAC, with her references to the underlying documents including now both the Swedish and the Spanish texts of the treaty, her investigation of the legal issues at stake and her recommendations for the solution of the case  was missing! Upon request of the whereabouts of this material it was first reported that it was not to be found! Nor was it electronically downloaded. However, after some further confusion, a copy belonging to the chairman of the Court was finally discovered and included into the file. And there, in the margin of the Spanish text, next to the term “rédito de fuente en el Peru”, the chairman of the Court had scribbled in pencil; “ej reavinst”, “not capital gains”.  Thus, seemingly, and only at this very late stage of the court proceedings, had one of the SAC judges spotted the Spanish version of article XVII § 2, reversing the whole outcome of the ARB ruling!

 

The SAC has thus concluded that the Spanish  term “rédito de fuente en el Perú” differs from Sweden’s “inkomst från inkomstkällor i Peru”.  The Spanish text represents in Swedish “avkastning från källa”   corresponding in English to “yield”, “proceeds”, “return” or “earnings”  (from sources in Peru).

 

Consequently, the interpretation problem now became focussed on the perceived divergence between the two authentic texts of the treaty. This kind of problem regarding plurilingual treaties is something that has been known to treaty making powers ever since treaties have been made. And the jurisprudence regarding this specific problem is massive. Ever since 1969 international law in these matters is embodied in the Vienna Convention on the Law of Treaties, ratified by Sweden in 1974. The rules of interpretation of treaties are to be found in articles 31-33 of the Convention. The specific problems relating to diverging texts of plurilingual treaties are of such importance and magnitude that they have been given a separate article, number 33.

 

The text of this article reads as follows:

 

Interpretation of treaties authenticated in two or more languages.

1.       When a treaty has been authenticated in two or more languages, the text is equally authoritative in each language, unless the treaty provides or the parties agree that, in case of a divergence, a particular text should prevail.
2.       A version of the treaty in a language other than one of those in which the text was authenticated shall be considered an authentic text only if the treaty so provides or the parties so agree.
3.       The terms of the treaty are presumed to have the same meaning in each authentic text.
4.       Except where a particular text prevails in accordance with paragraph 1, when a comparison of the authentic texts discloses a difference of meaning, which the application of articles 31 and 32 does not remove, the meaning which best reconciles the texts, having regard to the object and purpose of the treaty, shall be adopted.

 

 

The  interpretation problems deriving from diverging treaty texts have also been addressed by numerous international courts and tribunals both before and after the Vienna Convention entered into force.  International doctrine in this specific field of treaty law is also very prodigious.

 

In Sweden there are certain official documents covering also the problems of interpretation of plurilingual treaties, notably Government report “Internationella överenskommelser och svensk rätt” (International agreements and Swedish Law) (SOU 1974:100) and a Foreign Office report from the proceedings of the conference at Vienna in 1968-69. (UD 17 Mars 1970).

 

In Swedish case law, and, specifically regarding the interpretation of  tax treaties, the SAC has, as far as is known, only once touched briefly upon the issue of plurilingual treaties namely in RÅ 1987 ref. 162. More  hereinafter.

 

Important to note, however, is that the SAC on numerous occasions have revealed their strong adherence to the interpretation rules of the Vienna Convention. In RÅ 1996 ref. 84 for instance, the Court even took the trouble of explaining at length the contents of these rules in the Convention, firmly entrenching them as the foremost instrument for tax treaty interpretation.

 

This is what the SAC wrote in this case:

The interpretation of tax treaties shall aim at determining the common intentions of the contracting parties. This determination of the common intentions shall be done according to the methods and means advised under articles 31-33 of the Vienna Convention on the Law of Treaties (SÖ 1975:1). While this applies primarily to the relation between the contracting states, the same principles, as follows from case RÅ 1987 ref. 162,may also be adopted where interpretation takes place in a dispute between the tax payer and the tax authorities[5]. This means that the interpretation according to article 31 primarily shall be based on the text of the treaty according to the ordinary meaning of the terms, but also taking into consideration the context in the light of the object and purpose of the treaty (paragraph 1 of article 31). The “context” shall also comprise, in addition to the treaty text in its entirety, certain agreements and documents, which, at the conclusion of the treaty, have been made or drawn up between the parties (paragraph 2). In addition to ”the context” there shall be taken into account any subsequent agreements or practise which establishes the common understanding of the interpretation and also any relevant rules of international law applicable in the relations between the parties (paragraph 3). A term shall be given a special meaning if it can be established that the parties so intended (paragraph 4). Under article 32 recourse may be had to certain so called supplementary means of interpretation such as the preparatory works of the treaty and the circumstances of its conclusion, in order to confirm the meaning of the interpretation or to determine the meaning of the treaty when an interpretation according to article 31 leaves the meaning unclear or leads to a result that is manifestly absurd or unreasonable. In addition to these general guidelines, particular importance, in the field of international taxation, should be given to the OECD Model Treaty and the commentaries thereto that have been worked out by that organisation. If  a tax treaty or a provision in such a treaty has been worked out in accordance with the Model Treaty, it should normally be presumed that the contracting parties have intended to achieve a result corresponding to the recommendations of the OECD (see for instance RÅ 1987 ref. 158 and RÅ 1995 not.68 regarding the significance of the Model Treaty and its commentaries).

 

 

Scholarly work in Sweden in the field of interpretation of plurilingual treaties in general and tax treaties in particular is more or less nonexistent. The only work known regarding plurilingual tax treaties are the few pages written by this author in a 1988 report for the periodical issued by the Institute of Foreign Law  (Meddelande från IUR nr 10/1988, Om tolkning av dubbelbeskattningsavtal – Del I. sid. 23-29). This report, however, seems to be unknown to the SAC, at least it is not listed in the source materials submitted to the Court in this Sweden-Peru case.

 

The following, translated into English, is what I wrote regarding the interpretation of plurilingual treaties back in 1988:

 

 

_____________________________________

“For many centuries latin was the only written language in Western Europe. In step with the expansion of the great European nations, especially France under the resplendent court and victorious armies of Louis XIV in the 17th century, French emerged as the foremost diplomatic language. As of the late 19th century English gradually took over the leading role in the intercourse between nations.(The peace treaty of Versailles in 1919 was authenticated in both French and English, but of the five treaties that were made after the second world war only the treaty with Italy included an authentic French text.)

It is obvious that treaty interpretation based on the authentic method, i.e. an interpretation of “the ordinary meaning to be given to the terms”, will run into further problems if the text is authenticated in two or more languages. Therefore, it is not surprising that the Vienna convention has devoted a special article to this problem.

 

Also this limited area of  treaty interpretation has been the object of substantial analysis and scientific investigation. From the time preceding the Vienna Convention a good example is “The Interpretation of Plurilingual Treaties by International Courts and Tribunals” by Jean Hardy (37 Brit. Y.B. Int’l L. 72). With regard to the Vienna Convention there is also an excellent study by the Danish scholar Peter Germer: “Interpretation of Plurilingual Treaties:A Study of Article 33 of the Vienna Convention on the Law of Treaties” (1970 Harvard Int. Law Journal, 400).

 

Paragraphs 1 and 2 of article 33, regarding specific agreements that certain language version shall prevail, require no comments. The main rule regarding treaties that have been authenticated in two or more languages is, unless otherwise determined, that they shall be equally authoritative. In this matter the Swedish Foreign Office report submitted by the Swedish delegation from the Vienna Conference (H. Blix, H. Eek and R. Reuterswärd,, UD 17 Mars 1970) reads as follows:

“Even if a treaty has been drawn up in several languages – and thus allows a comparison between its texts – it remains a single instrument with uniform rules (a single set of terms), the interpretation of which is governed by articles 27-28 (article 31-32 of the final numbering of the treaty). When interpreting a plurilingual treaty the coherence of its terminology is of basic importance. This is maintained by the principle of the equality between authentic texts and that they shall be applied in the light of the presumption that the provisions are intended to have the same meaning in each text. This presumption requires that all efforts be made to find a common meaning of the texts. The fact that a treaty is plurilingual, does not allow the interpreter summarily to give preference to one text before the other and to neglect the means of assistance that exist, based on the purpose of the treaty, to remove existing discrepancies. Such means of assistance are for instance the preparatory works of the treaty and other circumstances at its origin as well as subsequent practise.
Paragraph 3 therefore concludes that the terms of the treaty shall be presumed to have the same meaning in every authentic text, and that, unless the contracting parties have agreed to give preference to a certain text when the authentic texts are different, one should choose the meaning which as far as possible reconciles them.

 

The Commission found no justification to include a general presumption for a restrictive interpretation in cases where plurilingual treaties gave rise to a wider or more restrictive meaning. Nor was it found motivated to establish a presumption in favour of the text that had the clearest text nor the language version in which the treaty had been negotiated.”

 

The common denominator when interpreting plurilingual treaties giving rise to ambiguities is thus always to try to reconcile the texts, not to give priority to any one of them before the other(s). The presumption of the uniformity of the text applies.


Germer concludes that because paragraph 4 of article 33 refers to the two preceding paragraphs the authors of the convention have expressed the opinion that (page 414):

 

“the very first rule for the interpreter of a plurilingual treaty confronted with a divergence between the different language versions is to look for the meaning intended by the parties to be expressed in the treaty by applying the standard rules for the interpretation of treaties. This means that the starting point of interpretation should be the ordinary meaning to be given to the terms of the treaty in its context and in the light of its object and purpose.”

As an example of a plurilingual treaty where the interpretation focused on its context, Germer refers to a case by the International Court of Justice in the Hague regarding the competence of the International Labour Organisation (ILO) visavi the Versailles treaty on working conditions for people working in the agricultural sector. Due to the French understanding of the term “industrie” and “industrielle” as being restricted to the fields of “trade” and” production” it was suggested that the ILO was not competent with regard to issues on agriculture. Any such restriction however was, according to the Court, not to apply to the English version of the term “industrial”. Seen in its context and considering the term’s appearance in article XIII of the treaty it was determined that it applied also to agricultural activities. “The context was the final test” argued the Court.

 

An interpretation of a plurilingual treaty taking into account its object and purpose was exemplified by the Wemhoff case, adjudicated by the Court of Human Rights in 1968. Karl-Heinz Wemhoff, a German citizen, had charged West Germany for violation of article 5 paragraph 3 in the Convention for the Protection of Human Rights. The article refers to the right by a detained or arrested person, according to the English language version
   

   “to be entitled to trial within a reasonable time”.

The French text reads:

 

    “a le droit d’être jugée dans un délais raisonnable ».

 

The question arose whether the reasonable time during which an imprisonment could continue should be considered have come to an end at the time of the commencement of the court proceedings, which was considered to be the case according to the English text, or, which the French text determined, at the point in time when the court proceedings were terminated.  The Court rejected the restrictive interpretation following from the English text and pronounced that:
 
“being confronted with a divergence between two equally authentic texts of a law-making treaty, it had to seek the interpretation that was most appropriate in order to realize the aim and achieve the object of the treaty, not that which would restrict to the greatest possible degree the obligations undertaken by the parties to the Convention.”

Article 33, as aforementioned, starts out from the presumption that each authentic text has the same meaning. If any doubt or obscurity occurs, one should attempt, at any cost, to reconcile the texts. During the drafting of the Convention by the International Law Commission (ILC), Alfred Verdross put forth the suggestion, when a reconciliation of the texts was not possible, to include a provision giving preference to the language version according to which the treaty had first been drawn up. However, Sir Humphrey (Waldoch), the rapporteur of the ILC, at the 874th meeting of the ILC, categorically rejected this suggestion, which had also been done at previous meetings. If a reconciliation was not possible, an interpretation should be made in the light of all relevant circumstances. A provision of preference for the original text was also unsuitable as the reason for the interpretation problem could have its root in that very text.

 

Germer too, is very firm on this point (page 418):

 

“It is sometimes asserted that in case of divergence between different authentic language versions of a treaty the text in which the treaty was first drafted is decisive. Under the Vienna Convention, however, this principle cannot be considered a valid rule. Once it is established by explicit stipulation or otherwise that the different language versions of a treaty are authentic, the interpreter is bound to give effect to the principle of equality of texts, and it would be a clear violation of that principle if the interpretation considered the original version to be superior to the other authentic texts. An examination of the preparatory work of a treaty and the circumstances of its conclusion may, however, display the causes of a divergence between the different language versions and thus help to establish the meaning intended by the parties to be attached to the provision in question.”

Sandels too, page 102,[6] categorically dismisses the idea  that preference should be given to the language that was used at the treaty negotiations.

 

At the Vienna conference a discussion arose regarding the reconciliation that the ILC draft had suggested regarding the interpretation of plurilingual treaties. The American delegation declared that the ILC’s recipe invited a kind of compromise without determining, however, the basis for how this compromise should be reached. The US delegate also maintained that a reconciliation of texts in many cases was impossible. The Vienna conference finally made an amendment to paragraph 4 of article 33, adding that the reconciliation should be carried out in the light of the object and purpose of the treaty.

This seemingly implies that the object and purpose of the treaty should be considered twice with regard to the interpretation of plurilingual treaties, first, says Germer (page 425),

”as a means by which a difference of meaning disclosed by the comparison of texts may be removed, and thereafter as a means by which language texts disclosing an “irremovable” difference of meaning may be reconciled in some way or other. This distinction is artificial. The problems of interpretation that are related to plurilingual treaties present no essential difference from the problems of interpretation that are raised by treaties which only have one authentic text. It is always the duty of the interpreter to look for the meaning which the parties intended to be attached to the terms of the treaty; the same principles of interpretation apply whether a divergence arises from the plurilingual form of a treaty or is found to exist between different provisions of a treaty established in only one language”.

A similar view was put forward by the General Advocate at the time, Maurice Lagrange, when, in his plea to the Court of the European Community in the Bosch Case (No 13-61, 8 Recueil de la Jurisprudence de la Cour 89, 1962), he argued regarding the four authentic languages of the Rome Treaty:

“Or, vous le savez, les quatre langues font foi, ce qui revient très exactement à dire qu’aucune d’elles ne fait foi…

Dans ces conditions, il faut se référer, tout comme on le fait en cas d’obscurité ou de contradiction dans l’interpretation des textes internes, au « contexte » ou à  « l’esprit du texte. »

The expression « l’esprit du texte », the spirit of the text, is very appropriate. This is the beacon to be followed in all efforts of treaty interpretation. This is what is meant by interpretation in good faith.

__________________________________________

 

 

 

Diverging treaty texts and article II §2 (the lex fori rule).

 

The initial statement by the SAC that both the Swedish and the Spanish texts are authoritative for interpretation purposes is of course correct. The following conclusion, however, that, by reference to article II § 2 of the treaty – the so called lex fori rule – the Swedish text shall have precedence when the treaty is applied in Sweden but that in cases of uncertainty the Spanish text, too, shall be considered is not.  The reason herefor is that the lex fori rule is an instrument for determining the meaning of terms that have not been defined in the treaty which may lead to the adoption of the internal law meaning of the term of the country applying the treaty and this has nothing to do with the interpretation problems stemming from divergent language versions. This is a different problem that shall be resolved by application of the above cited rules in article 33 of the Vienna Convention. And, in contrast to the conclusion of the SAC,  the fundamental principle of article 33 and the starting point for its application is the equality of the texts, not that any of them shall take precedence. Nor, of course, does article 33 of the Vienna Convention refer to any internal law meaning of the term being interpreted.

 

However, one must also keep in mind that the problem facing the interpreter in this case is indeed  also to determine the meaning of an undefined term in the treaty, namely the term “income from a source”. The interpretation task of the case is thus twofold; a) to solve, according to article 33 of the Vienna Convention, the problem of the diverging language meanings and b) to determine, according to article II § 2 of the treaty the meaning of an undefined term.  It is, however, a difficult task to separate these two sets of rules because it is likely even if  not intentional that the different languages of the contracting state will tend to reflect also these states’ domestic law meanings. But the fact that the SAC has based its interpretation approach only on the lex fori rule of the treaty nourishes the suspicion that the Court has not contemplated article 33 of the Vienna Convention at all.

 

In the ‘dilemma’ arising from the fact that both the lex fori rule of the tax treaty and article 33 of the Vienna Convention on plurilingual treaties come into play in this case, it is also well worth citing again  Peter Germer when, winding up his article, he concludes that “ the problems of interpretation that are related to plurilingual treaties present no essential difference from the problems of interpretation that are raised by treaties which only have one authentic text. It is always the duty of the interpreter to look for the meaning which the parties intended to be attached to the terms of the treaty; the same principles of interpretation apply whether a divergence arises from the plurilingual form of a treaty or is found to exist between different provisions of a treaty established in only one language”.

 

The SAC itself, which will be further explored below, has also established, and very eloquently for that matter, that the overriding purpose of treaty interpretation is the determination of the intention of the contracting parties.  So, the question to be answered in this case is if it was it the intention of Sweden and Peru to exclude capital gains from the scope of article XVII § 2 or not.

 

 

The meaning of  “income” and “from a source”.

 

Article 33 first advises the interpreter to attempt to remove the differences of meaning in a plurilingual treaty by applying the general rule of interpretation in article 31 and then the supplementary means of interpretation in article 32.

 

The main conclusion of article 31 is that one should adopt the ordinary meaning of the pertinent treaty. The following will represent a discussion to analyse the SAC’s interpretation  of  “income from sources (in Peru)” from this standpoint. Article 31 also proclaims that the ordinary meaning shall be considered “in its context and in the light of its object and purpose”. This will be more closely considered in a later section of this analysis.

 

a) “Income”
 
In what seems to be an attempt to apply the ordinary meaning according to the Vienna Convention to the term “income” in a tax treaty the ARB, as mentioned above, had no problems in establishing that this term comprised also capital gains.

 

A first reaction to the ruling given by the SAC, that came to the opposite conclusion, is their statement that  “income from sources in Peru” in its Swedish language version is “not conclusive” for determining whether it comprises capital gains. Only in the light of the Spanish version is it made clear that this is not the case. A supporting factor in this regard according to the SAC is also that Peru, as evidenced from the taxes covered article, makes a distinction in its internal law between income (rentas) and gains (ganancias).

 

The statement by the SAC that the Swedish language version of the treaty regarding the scope of the term “income” does not conclusively answer the question  whether or not capital gains are included therein is quite surprising because it completely ignores the explicit wording to the contrary in the text itself of article 2 subparagraph 2 of the OECD Model Tax Convention (already in its version of 1963 which thus existed when this treaty was negotiated ) and  is extensively discussed both in its Commentaries and in international tax doctrine. And what is quite ironic is that this conclusion  by the OECD is something that actually springs also from the taxes covered article itself, i.e. the article to which the SAC refers for its reason leading to  its contrary conclusion! Or in other words; where, on the one hand, the conclusion by the SAC that the term “income” excludes capital gains takes its justification from a distinction in the nature of the Peruvian taxes enumerated in the taxes covered-article, the authors of the OECD Model treaty, on the other hand, with reference to that same article thoroughly dismiss this approach.

 

According to the wording of article 2 subparagraph 2 of the OECD Model Treaty“there shall be regarded as taxes on income…all taxes imposed on total income…or on elements of  income…including taxes on gains from the alienation of movable or immovable property (emphasis added), taxes on total amounts of wages or salaries paid by enterprises, as well as taxes on capital appreciation”.

 

Therefore, and regardless of their nature or description under the domestic laws of the contracting parties,  the taxes enumerated in and covered by article 2 shall, for the purposes of application of the treaty, always be considered as taxes on “income”.

 

Klaus Vogel on “Double Taxation Conventions” (Kluwer 1991) at page 89 also comes out very strongly on this point. He concludes that “the context of the distributive rules of Arts. 6 to 22 MC reveals all the items which, according to MC, may be subjected to income tax. MC consequently presupposes that despite all the differences in the details of various arrangements, there is at international level a basic common understanding of what ‘income’, the French term ‘revenue’ , and the German term ‘Einkommen’ mean….The positive definitions of the term ‘income’ in national income tax legislation usually are much narrower than the widest of all definitions of the term. But that does not prevent the corresponding taxes from being subsumed under the treaty term ‘taxes on income’. It is also irrelevant whether the tax rate is applied to ‘income’ or ‘receipts’ or ‘yield’ or ‘profit’ or to any other less sophisticated yardstick as long as the aim is to cover income in the afore-mentioned widest sense, or essential elements thereof. Income also includes, as stressed by MC, gains from the alienation of movable or immovable property”.

 

It is quite common, especially in countries such as Peru which, at the time of the conclusion of the treaty, had a schedular tax system, that capital gains taxes are separately taxed from other types of income. This is also echoed by the preliminary remarks (paragraph 3) of the commentaries to article 13 on capital gains in the OECD Model Treaty regarding the variety of capital gains taxes in the internal laws of treaty making states together with a direct reference to article 2. Thus, according to the OECD Model Treaty commentaries of article 13 “It is understood that the Article must apply to all kinds of taxes levied by a Contracting State on capital gains. The wording of Article 2 is large enough to achieve this aim and to include also special taxes on capital gains.”

 

The list of situations where the term “income” must be given this all-encompassing meaning is very long. For instance, the avoidance of double taxation article (number 23 of the OECD Model) speaks only of “income” and it would be absurd to suggest that it would not provide for tax relief with regard to a double taxation of capital gains. When interpreting the full term “income from sources” for the purpose of determining residence under the treaty as it appears in article 4 subparagraph 1 of all OECD Model-based treaties this term must of course also cover capital gains. In article XVII § 3 of the Sweden-Peru treaty allowing Sweden to impose its graduated rate of tax, the text also mentions only “ income….exempted under this Agreement were included in the amount of total income…”   No serious treaty interpreter would suggest that this (exemption with progression) article were not applicable to an exempted capital gain just because Peru  taxes such gains under a special domestic tax law separately itemized in article I of the treaty. Nor, of course, would anyone dream of proposing that the Sweden-Peru treaty has not taken effect with regard to capital gains just because the relevant article (number XXI) only mentions income derived on or after 1 January of the calendar  year …. etc” Also, generally, tax treaties include an “other income” article (number 21 of the OECD Model Treaty) referring to income not previously mentioned in the treaty. Certainly such “income” must comprise also capital gains for interpretation purposes.

 

Moreover, the Peruvian capital gains tax mentioned in the treaty no longer exists. Instead, today, capital gains taxation is included in the general Peruvian income tax act (Ley del impuesto a la Renta). Considering in this context a) that tax treaties shall apply also to any identical or substantially similar taxes that are imposed after the date of signature of the treaty in addition to, or in place of, the existing taxes and b) that an ambulatory approach shall be adopted when interpreting the meaning of treaty terms, it would seem bizarre , in a situation like this, to suggest that the nature of an abolished tax regime shall be decisive for determining the meaning of treaty terms as they are understood today.

 

 

b) “From a source”

 

Many analysts exploring the question of the source of an income are inclined to get involved in quite theoretical and academic considerations of the meaning of this term usually cog-wheeled some way or the other to domestic rules or to general economic theory[7].

 

This, however, is immaterial when applying this term in a tax treaty where the determination of the source of the taxable items under the treaty is quite straightforward and based on principles recognised by the OECD Model Convention, see Klaus Vogel,  page 1017  of “Double Taxation Conventions” (Kluwer 1991). This principle, which sometimes - especially in treaties with Anglo-American States - is embodied directly in the treaty, e.g. article 22 subparagraph 4 of the Sweden-UK treaty, ordains that whenever an income or a gain may be taxed under the treaty by the contracting state that is not the state of residence of the tax payer, the income or gain is also considered to have its source or is deemed to arise in that (source) state. If the right to tax is assigned primarily to the state of residence this will be the determining  justification  herefor and the matter of the source of the income becomes  irrelevant (see above the dissenting ruling by Mr Virin in the ARB).

 

Consequently,  in this case,  as the gain shall be taxed only in Peru under article X and is derived by a person resident in Sweden this will mean that the gain for treaty purposes also shall be deemed to arise or have its source in  Peru. For this reason the conclusion reached by  Wingren and Brydolf  in their dissenting opinion (see above) that the gain was not sourced in Peru must be rejected.

 

The interpretation of the term “income from a source” in the (1968) tax treaty between Sweden and the UK has been discussed extensively in RÅ 1987 ref. 162, see further below,  regarding the subject to remittance rule of that treaty in its application to a case where a UK resident did not remit the gains of the sale of Swedish shares to the UK. In this treaty there was no divergence between the two languages of the treaty . Instead the interpretation issue focused on article 3 § 2, the lex fori rule, regarding the extent to which the domestic (Swedish) law meaning of the term “income from a source”, should be adopted .  In its investigation hereof the SAC, convinced that this represented the common intention of the contracting states, ironically enough came to the conclusion that the term  should be given the meaning it had under the laws of the other country, the UK, which did not comprise capital gains, not under the laws prevailing in Sweden that was the state applying the treaty.

 

As gleaned from the above analysis the term “income from a source” in whatever context it appears should be given the meaning it has neither under Swedish nor UK nor Peruvian nor any other domestic tax law for that matter. It should be given the ordinary and well established meaning it has under ‘tax treaty law’ or, more appropriately, under the principles that exist and the experiences that can be drawn from tax treaty application and interpretation in general and in harmony with the OECD Model Convention in particular. “Income from a source”, as just mentioned, is a common term that can be found in each and every tax treaty. And there are no devious minds plotting to limit its application!  On the contrary, and responding to the very first dictum of the  Vienna Convention’s interpretation rules that one shall interpret international agreements “in good faith”, this term, if nothing to the contrary has been explicitly and clearly  determined by the contracting parties, should be understood in its widest sense, serving to its fullest the purpose of preventing both double taxation and double non-taxation.

 

 

The meaning of “réditos de fuente” in the Spanish text of the treaty.

 

The bottom line conclusion of the SAC ruling is that the Spanish term “rédito” in article XVII § 2 has a more limited scope than “inkomst” (income) in Swedish applying, as it seems, only to  taxable items that generate a current income from invested capital such as a dividend but not to the situation where an investment is alienated. And this according to the SAC is made especially clear by the existence in Peruvian tax law of a separate  regime for taxing “rentas”  and another  for imposing tax on “ganancias”, both  itemized in the taxes covered article in the treaty. The treaty term “réditos de fuente” in article XVII § 2 can thus be applied  only to such taxable items that are taxed under the income tax law (“impuestos sobre la renta “) but not to taxable items taxed under a capital gains tax law such as “impuesto a las ganancias (de capital provenientes de la transferencia de immuebles)”. The ruling thus focuses its attention on a distinction between on the one hand the Spanish meaning of a “renta” as it appears in the Peruvian income tax law  and on the other hand “ganancias” in the capital gains tax law.

 

As already mentioned Peru, at the time of the conclusion of the treaty, had a schedular tax system leading to a structural separation of its income tax and capital gains tax acts. This, however, is something that follows from general tax policy considerations, prevalent in less developed economies , and not from any theoretical perceptions of any differences between the nature of these two taxable items.

 

Without claiming any definite authority of the finer nuances or distinctions of the Spanish language, a webb-search of the “Diccionario de Sinónimos y Antónimos” at www.wordreference.com indicates, however, that the Spanish terms “renta”, “ganancia” and “réditos” are all synonymes.  For instance, in the present Sweden-Argentina treaty the Argentine income tax act is called in Spanish not “impuesto sobre la renta” as in the Sweden-Peru treaty but “impuesto a las ganancias”.  Consequently, there seems to be no such firm  distinction linguistically or otherwise between  the term “rentas” and “ganancias” nor any absolute language barrier to prevent the application of  “réditos” to both of these terms as suggested by the SAC. Besides, according to contemporary economic theory as reported at the Swedish website  www.finansportalen.se, priding itself of having the best economic word-list in Sweden on the Internet, the term “avkastning” (réditos),  is defined as “Income resulting from an investment. With regard to (company) shares it can arise from either dividends or from an increase of the share value”.

 

Also, according to reports by Peruvian tax expertise the term “réditos”  comprises for tax purposes both (current) income derived from business activities and capital investments on the one hand but also, on the other hand capital gains derived from the alienation of an investment.

 

Therefor, even if the authentic text of the treaty had been in Spanish only, it is not self-evident that the term “réditos de fuente” should be given a meaning excluding capital gains.

 

In an attempt to reconcile the relationship between  the different Spanish terms of the taxable items of the treaty and in observance of the general meaning of the term “avkastning” proposed by the SAC as the proper translation of “rédito”, it is suggested that the ordinary meaning under article 31 of the Vienna Convention of the term “rédito de fuente” in its context and in the light of its object and purpose in article XVII § 2 (see below) can indeed be applied also to a capital gain on the sale of shares and thus consonant with the Swedish text “inkomst från inkomstkälla” (income from a source).

 

Also telling is the English translation of the Sweden-Peru treaty published by the much  respected International Bureau of Fiscal Documentation (IBFD) in the Netherlands. To no surprise (see also above), the IBFD´s version of article XVII §2 based on both the Swedish and the Spanish texts also reads  “income from sources” corresponding to the Swedish text. No attention has been paid to the Peruvian text. The IBDF’s choice of the term “income from sources” is probably due to the fact, as already noted above, that it is a very common and ordinary expression in tax treaties, but it could of course also be because they have found that the Spanish term does not constitute a linguistic divergence. The English translations made by the IBFD, which are the result of work which has been carried out after the publication of the authentic texts, are of course no part of the preparatory work of the treaty nor formally authoritative for interpretation purposes, but considering the authority and experience of the IBFD in these matters and that this organisation is completely independent and un-biased, their English texts should nevertheless not be ignored in cases where a divergence between the authentic texts arises.

 

 

Supplementary means of interpretation according to article 32  of the Vienna Convention.

 

The next step in the process of interpreting plurilingual treaties under article 33 of the Vienna Convention is to determine if and to what extent article 32 can be helpful in removing any perceived differences of the authentic texts.

 

According to this article recourse is to be had to supplementary means of interpretation, including the preparatory work of the treaty in order
a) to confirm the meaning resulting from the application of article 31 and or
b) to determine the meaning when the interpretation according to article 31 leaves the meaning ambiguous or obscure or leads to a result which is manifestly absurd or unreasonable.

 

The preparatory works of tax treaties are very few. However, the negotiation of these treaties made by Sweden are always conducted in English and will thus always result in an initialled draft in that language[8]. The Swedish Government Bill (prop. 1967:26) of the Sweden-Peru treaty explicitly mentions the existence of this initialled English version. This draft of the treaty is clearly a part of the preparatory work thereof  and one would imagine that every responsible interpreter, facing a plurilingual treaty with a divergence in its  authentic texts, would jump at the opportunity to examine such a third English text.  However, and very regretfully, this has been neglected by the SAC.

 

Upon request, and after some difficulties, the English initialled draft of the Sweden-Peru treaty was retrieved from the Swedish National Archives, (Riksarkivet).[9] It contains interesting information not only regarding the divergence between the two texts of article XVII § 2 but it also casts light, as shall be duly reported, on the negotiation process which is of  interest in general to the understanding of  the final text(s).

 

And how then does article XVII § 2 read in this English text drafted by the negotiating teams, invariably consisting of very experienced international tax specialists and high-ranking government officials of  the contracting states and initialled (page by page) by their chairmen?  Does it read something like “yield/proceeds/return or earnings from sources in Peru” supporting the Spanish version of excluding capital gains? Of course it does not. Such an expression does simply not exist in the normal tax treaty vocabulary. No, and to no surprise, the English text of article XVII § 2 in the draft reads “income from sources in Peru”. This, as already mentioned above, is the ordinary and established term that will be commonly found in almost all tax treaties. Consequently, if the SAC had attempted to confirm its interpretation by taking recourse to the preparatory work of the treaty and the circumstances of its conclusion according to article 32 of the Vienna Convention they would have found no support for their conclusion giving primacy to the Spanish text. Quite the opposite, the initialled draft clearly confirms that the ordinary meaning that should be given to the term “income fro sources in Peru” is that which follows from the Swedish text. Under all circumstances, the conclusion to be drawn herefrom is that there was a consensus between the negotiating parties at the time of the initialling of the draft treaty and that the common intention then was to give the subject to tax rule in article XVII § 2 as wide a scope as possible encompassing all taxable items that could potentially go untaxed altogether. 
 
The Sweden-Peru draft was initialled  upon the conclusion of the (face to face) negotiations which took place in Lima in early 1964. During the subsequent exchange of letters two material changes of the initialled text took place, one regarding article XVII § 2, the subject to tax article and the other regarding article X the capital gains article, i.e. both of the articles which are at the centre of our interpretation efforts . As shall be demonstrated, no effort was undertaken during this phase of the negotiations to limit the scope of article XVII, which, if it had occurred, would of course have been reflected in the Swedish text also.
 
A further point worth mentioning regarding the initialled draft  is that the Peruvian taxes to be covered by the treaty  in article I were not at all mentioned therein. This was left open for further consideration. Thus, when negotiating and initialling a tax treaty for the avoidance of double taxation  on “income” (renta) as determined in the title of the treaty, there was no reason to believe that, at this point in time, the subsequent inclusion of the Peruvian  capital gains tax (ganancias) was intended to give rise to any limitations of the scope of article XVII § 2.

 

The translation of a treaty  into its authentic texts is always carried out separately by the contracting states after a final agreement is reached in what in this case was the English text. As established above the initialled text of article XVII was  “income from sources in Peru” which comprises also capital gains. It is therefore obvious that the problem of the diverging authentic texts, as perceived by the SAC, must have its root in the translation work. The Swedes, accustomed to this term from many previous treaties and suspecting no problems herewith, translated the English text of article XVII § 2 into “inkomst från inkomstkällor i Peru”, which of course gives rise to no problems. The Peruvians translated the English text into “réditos de fuente en el Peru”. Hereafter, which is the standard procedure, the two texts are exchanged for mutual scrutiny. Thereupon in Sweden the texts were submitted to the Stockholm Administrative Court of Appeal who found no differences between the two texts.  One must have a very conspiratorial mind to suggest that either of the contracting states at this stage of the treaty making process were engaged in any purposive effort of giving “income from a source” any other (limited) meaning. Rather, if “réditos” is accepted as a linguistic limitation, one is tempted to suspect that it is nothing but an innocent and - most importantly - unintentional translation error which, possibly due to the fact that this was the very first tax treaty ever made by Peru, resulted from inexperience with tax treaty matters.[10] 

 

The preceding investigation of the original English version drawn up at the negotiations and the subsequent events of the treaty making process is a good  example of an application of the rules of article 32 in the Vienna Convention according to which, for interpretation purposes, “recourse may be had to the circumstances of the conclusion of the treaty.“ An investigation of this kind also corresponds to the SAC’s own recommendation for the procedure to be followed regarding the interpretation of undefined treaty terms according to the lex fori rule, that one shall examine “the circumstances of the introduction of the treaty and its historical context”, (see below). It gives a much deeper understanding of the intentions of the contracting states than that which follows from a superficial textual analysis of, in this case, a perceived divergence of the treaty texts.

 

Under article 32 of the Vienna Convention recourse to preparatory materials  shall  be had not only to confirm the meaning reached according to the interpretation rules in article 31 but also in order to  determine whether this meaning  (a) appears ambiguous or obscure; or (b) leads to a result which is manifestly absurd or unreasonable.

 

It is noteworthy, not to say surprising, that the authors of the Vienna Convention have found that also a careful and conscientious interpretation under article 31 can indeed lead to a result which is manifestly absurd or unreasonable. The term “manifestly” even seems to imply a kind of qualified level of absurdity or unreasonableness.

 

In his pre-Vienna Convention study from 1948, “International Law: A treatise”, Lassa Oppenheim also concluded that treaties must be interpreted in their “reasonable, in contradiction to their literal sense” and , he went on: “If, therefore, the meaning of a stipulation is ambiguous, the reasonable meaning is to be preferred to the unreasonable, the more reasonable to the less reasonable, the adequate meaning to the meaning not adequate for the purpose of the treaty, the consistent meaning to the meaning inconsistent with generally recognised principles of International Law and with previous treaty obligations towards third States.” A favourite example hereof, quoted also by many other scholars, was when France, in the interest of Great Britain and in compliance with article 9 of the Peace Treaty of Utrecht in 1713, demolished the harbour and all its fortifications at Dunkirk, but at the same time started to construct an even larger naval base at Mardyck just a couple of miles down the Channel coast! (Great Britain, however, successfully protested that France hereby was in breach of “the reasonable if not the literal sense of the Peace at Utrecht.”)

 

So article 32  instructs the interpreter to contemplate the result of his interpretation. The provision can therefore be considered a “warning bell” which, if it chimes, should spur him to reconsider his interpretation effort. It is up to each and everybody’s judgment to determine if the SAC ruling is “manifestly absurd or unreasonable” but it is definitely in clear violation of the object and purpose of article XVII § 2 of the Sweden-Peru treaty and it disregards the functional approach to the structural composition or design of tax treaties. This will be explored in the following sections.

 

 

The object and purpose of the treaty.

 

As earlier mentioned article 33 of the Vienna Convention advises the interpreter first to solve the problem of divergent treaty languages by applying the normal rules of interpretation under articles 31 and 32. An important element in this process according to article 31is to determine the ordinary meaning of the terms of the treaty in their context and “in the light of their object and purpose”. If this does not remove the divergence one should adopt the meaning, which best reconciles the texts with “the object and purpose of the treaty”.  Interesting to note is thus that the object and purpose of the treaty comes into play under both these procedures emphasising the importance of this element in treaty interpretation in general.

 

The fact that one should interpret the treaty “in the light of” its object and purpose further suggests that this should continuously remain ‘the backdrop’ to the interpretation process. The determination of the object and purpose is something which should therefore be done at the very start of the interpretation. If subsequently the final result of the interpretation is incompatible with the object and purpose of the provision under scrutiny then one must re-examine one’s interpretation and investigate if something has been overlooked. One must thus make sure that all the options indicated in subparagraph 2 of article 31 in the Vienna  Convention have been considered and ask oneself if, according to article 32, there are any supplementary means of interpretation that should be observed such as the preparatory works of the treaty or any other circumstances prevailing at the time of its conclusion.  

 

As can be gleaned directly from the SAC ruling, the purpose of the interpretation process is to establish the common intention of the contracting parties. It is suggested that this approach means the same thing as saying that one should interpret the treaty in accordance with its object and purpose. They both represent the two sides of the same coin. In his recent dissertation “Taxation of Cross-border partnerships” (page 41) Jesper Barenfeldt“ convincingly argues that “the object and purpose of a treaty is a reference to the intention of the parties and that the object and purpose of a treaty can hardly be anything but the intended purpose of the parties”.

 

The intention of the contracting parties thus represents ‘the holy grail’ of treaty interpretation and this shall be sought by faithful and  continuous adherence to the object and purpose of the treaty.

 

So what then is the object and purpose of the subject to tax provision in article XVII § 2 of the Sweden-Peru treaty? The answer hereto is perfectly clear; it is to prevent double non-taxation. It is a well-known  fact that tax treaties, especially when the exemption method for eliminating double taxation is used, bears a high risk of giving rise to double non-taxation when, for some reason, the income is not taxed in the source state and therefore it is common practise to include subject to tax provisions in such treaties.[11] Apart from the very special situations regarding tax sparing provisions (see further below) or in such a case which follows for instance from article 26 subparagraph 3 in the Nordic treaty[12] , states that conclude treaties for the avoidance of double taxation never have the intention to allow double non-taxation or complete tax exemption. At IFA’s congress in 2004 on double non-taxation professor Hugh Ault at OECD’s Centre for Tax Policy and Administration and one of the members on the discussion panel, reminded the delegates of the 1927 League of Nation report that every effort should be made not only to avoid double taxation but also to ensure that all income be taxed once, and only once. The starting point of every tax treaty making process is of course that there is a tax claim in at least one of the contracting states. This is always reflected in the taxes covered article. And there is no intention that this single tax claim shall be wiped out entirely by the treaty.  The only and quite self evident situation where double non-taxation is  normal is where neither of the contracting states has a tax claim under its domestic tax laws.

 

This attitude to double non-taxation is reflected also in the Swedish Government Bill introducing the Nordic Multilateral Tax Treaty to Parliament (see pages 91-94 of Riksskatteverkets publication of that treaty 2000) where the government in a long statement regarding its general views of treaty interpretation addresses also the topic of the purpose of the treaties: “Moreover, the lodestar for the interpretation of tax treaties shall be to respect the purpose thereof i.e. that double taxation is avoided and that tax evasion and double non-taxation is prevented. The starting point of the interpretation process is thus that if Sweden according to its domestic laws may impose taxation of certain income or property and has not explicitly given up its right to such taxation then (then that income and property) should be taxed in Sweden”.

 

Moreover, with regard particularly to double non-taxation, one should also take into account the very special situation that follows from making a treaty with a country like Peru, which at the time of the conclusion of the treaty, (like several other states in South America at the time), operated a territorial tax system, meaning that all foreign or third state sourced income earned by Peruvian residents was exempt from tax. This fact is also mentioned in the Government Bill of the treaty. Such a tax system has the effect that if one makes a tax treaty with the country in question without a subject to tax clause and simultaneously adopts the exemption method for eliminating double tax, all third state income will automatically go untaxed. Consequently, when making a treaty with a country with a territorial tax system, the risk of double non-taxation is especially high, requiring always the inclusion of a general subject to tax rule. Sweden, like most other states, has been very observant in this matter always – there is not a single exception hereto - including subject to tax rules in treaties with countries operating territorial tax systems.[13] (Since 1993 Peru has switched to the world wide tax liability system granting a credit for foreign taxes to avoid double taxation.)

 

In her presentation of the case to the SAC Ms Wetterfors analysed the object and purpose of the treaty in order to establish the intentions of the contracting parties. According to the title of the treaty, she declared, the purpose thereof was to avoid double taxation, simultaneously pointing out that the purpose of preventing  tax avoidance was missing in the title. She noted that this latter purpose was missing also in the titles of the treaties with Argentina, Brazil and the U.S (of 1965) but was included in the title of the 1956 treaty with South Africa. The purpose, she finally concluded, of articles X and XVII § 2 of the Sweden-Peru treaty was likewise to avoid double taxation. No reference, however, was made to the purpose of the latter article to prevent double non-taxation.

 

It is true that the title (and the preamble) of international treaties can be a useful instrument for establishing the object and purpose thereof . This however shall not be understood that all the various objectives of a treaty must be included in its title to be recognised. An interesting illustration hereof is the fact that the OECD Tax Treaty Model  has chosen to refrain completely from mentioning any of its purposes in its title, simply referring to its function as a “Tax Convention on Income and on Capital”. A comparison of the preamble of the Sweden-Peru treaty with other treaties concluded by Sweden is thus of  little importance.

 

The purpose of preventing double taxation with regard to the vast majority of cases where Peru indeed taxes the income has already been determined, in this case in article X. Article XVII § 2 on the other hand is to take care of those cases where the income is untaxed in Peru. Or, in other words, article XVII § 2 is not an article to prevent double taxation, it is an article to impose taxation once in order to prevent double non-taxation. Therefore it must be considered negligent not to recognise the importance of this specific subject to tax element of article XVII §2. One must search for the object and purpose of the treaty not only in its title or preamble but also, as far as possible, in the provisions themselves and only after having considered the whole of the treaty is it possible to determine the full scope of its objects and purposes.

 

Nevertheless, and already during the proceedings  in the ARB, and evidenced especially by a document from the applicant companies, a discussion arose from the suggestion that article XVII indeed was not a subject to tax rule at all! This idea was subsequently endorsed also by Ms Wetterfors who concluded that “article XVII is not a subject to tax article as claimed by the tax authorities but that the treaty is a pure exemption-treaty rendering it irrelevant whether Peru exercises its taxing rights or not”.

 

This point of view, however, was rejected already by the ARB by their conclusion that the wording of  article XVII was quite clear and that any other interpretation would render it quite meaningless. Or, quite simply; a rule that explicitly stipulates a subject to tax condition is also a subject to tax provision. The nature of article XVII as a subject to tax rule is also confirmed by the SAC ruling, proving at the same time that the Court were perfectly aware of the fact that the purpose of the provision is to prevent double non-taxation.  Thus, firstly, the Court determines that the right of tax is assigned to Sweden “if Peru does not exercise” its (primary) right to impose tax. Secondly, in the last phrase of their ruling,  the Court concludes “that it is immaterial whether or not the gains have been effectively taxed in Peru.”

 

The language of the SAC in both these statements  reveals, however, a somewhat careless or imprecise approach to this problem disclosing a serious misunderstanding of the different meanings of  the just mentioned expressions.  The question whether Peru has or has not “exercised” its right of tax or whether or not the gains have been “effectively taxed” (in Peru)  is something which is completely different from the condition in article XVII that the income be“subject to tax” (in Peru). The first two situations are relevant when there is a possibility  impose tax under the laws of the state in question but where this for some reason has in fact not been effectively exercised [14], whereas on the other hand the subject to tax condition indicates that , in the absence of legislation, there is no such right to tax (in this case in Peru)[15]. As mentioned, for the purposes of the ruling itself, it was assumed that no taxing rights existed under Peruvian law with regard to the pertinent gains. In the citation above of Ms Wetterfors’ statement that article XVII § 2 is not a subject to tax rule the same confusion of these different tax conditions is revealed. For further study of the just mentioned distinctions, see the extensive discussion hereon in both the national and general reports of the 2004 IFA congress,  Cahiers de Droit Fiscal International , Volume 89a on “Double non-taxation”.

 

 

 

 

 

Double non-taxation.

 

The problems regarding double non-taxation have been extensively discussed in the last couple of years by the international tax community. Thus, it was the main subject of the congress held in 2004  by the International Fiscal Association (Cahiers de Droit Fiscal International, volume 89 a (Kluwer)) and discussed too by this author in a recent article, see www.skatter.se.

 

In many situations it is difficult to establish the object and purpose of specific tax treaty provisions. This, in particular, is the case regarding the distributive rules of the treaty (under chapter III of the OECD Model Treaty) dealing with the different classes of income. This is due to the reciprocal nature of these rules where the final outcome of which country shall have primary taxing rights will depend on which country is the state of residence of the tax payer and which state is the state of source of the income. Or, in other words, these articles merely provide the technical framework for the avoidance of double taxation and it is not possible to determine any specific intention by the contracting states in these matters for interpretation purposes. [16]

 

But, with regard to a subject to tax rule, the purpose thereof, which has already been explained, is quite clear ; namely the prevention of double non-taxation. At first glance article XVII § 2, however, gives, as it seems, a strange message in that it establishes that the residence state shall give exemption from tax where the other state has primary and sole taxing rights explicitly noted already in each and every distributive rule! Thus, article X for instance has already established that capital gains shall be taxed only in the state where the asset in question was situated at the time of sale etc. In such a situation a further general rule determining a tax exemption in the residence state seems superfluous. Why then is this exemption method nevertheless repeated in article XVII?  There answer hereto is twofold: Firstly  article XVII provides the  further qualification for the exemption already mentioned in the foregoing articles namely the subject to tax condition. Article X is thus the article that determines the treatment of gains that are subject to tax in Peru covering what is probably the vast majority of cases, whereas article XVII § 2 takes care of the situation where the gain has not been subjected to tax in Peru. This qualification could of course technically have been included in each and every ‘income’ article but such a repetitive method would be linguistically awkward and represents a technique which is simply not used in tax treaties. Subject to tax rules are thus always placed at the end of the treaty text,  naturally referring to all the preceding distributive rules of the treaty.  Secondly , and this is explicitly mentioned in section A of the Swedish general directives of the treaty with Peru, article XVII § 2 serves the purpose of determining that the exemption method shall apply to income (and gains) that have not been previously mentioned in the preceding articles of the treaty, including third state income.

 

An important observation of the SAC ruling is that also Swedish resident individuals may benefit from its double non-taxation consequences. Considering the Swedish tax system in general and Sweden’s treaty policy in particular, the suggestion that this was the result of a deliberate intention by the Swedish negotiators represents a serious mistrust of their competence! Such a sacrifice of  Swedish domestic tax claims has simply never occurred in the history of Swedish tax treaties. Therefore, a suggestion that our treaty negotiators would be chasing around the world intentionally putting in double exemptions of this nature in our treaties is, in the words of the Vienna Convention, truly “manifestly absurd and unreasonable”. 

 

Moreover, if in fact such a deliberate aberration had been made from Swedish tax treaty policy it would definitely have been mentioned in the Government Bill requesting  Parliament approval of the treaty. Anything else would have constituted a serious breach of duty by the Swedish Government. 

 

 

 


Structure of article XVII of the Sweden-Peru treaty

 

It is also necessary to investigate closer the ‘structure’ or the ‘building blocks’ of article XVII § 2 to fully understand its scope. This, together with the “systematic approach and  the function of the article in question”, is something that the SAC itself, (see again below regarding RÅ 1987 ref.162), has repeatedly underlined in earlier tax treaty interpretation cases. And here, again, the initialled draft of the Sweden-Peru treaty  provides interesting information!

 

As explicitly declared in the opening subordinate clause of article XVII § 2  – “subject to the provisions in article VIII…”(Där icke bestämmelserna I artikel VIII annat föranleda…)  -  royalties are exempted from its scope. Untaxed royalties in Peru shall thus be exempted from tax also by Sweden and consequently enjoy double non-taxation. This is perfectly intentional and a consequence of a deliberate desire to encourage the transfer of  Swedish know-how to Peru. At first glance, however, it seems completely unnecessary and contradictory to exclude royalties from article XVII as this type of income, in contrast to all the other types of income, may in fact be taxed under article VIII by Sweden (as the state of residence of the licensor) In such a  situation where Sweden effectively imposes tax, a double non-taxation will of course never occur. But the above mentioned desire to encourage a transfer of technology to Peru is instead accommodated by the tax sparing provision in article VIII § 4.  The reason, however, for excluding royalties from the subject to tax rule is  perfectly logical! It is there to cover the situation where a royalty income is attributable to a permanent establishment in Peru. Such a permanent establishment is taxed only in Peru under article III § 2 and exempted from tax in Sweden. If then in such a situation, Peru, for whatever reason, would not tax a royalty attributable to a permanent establishment it would be subject to tax in Sweden under article XVII if the exception therein regarding royalties had not been included.  The double non-tax situation which will thus occur according to the subordinate clause of the text reflects the intention by Sweden also in this permanent establishment context to grant special tax relief  to Swedish licensing  investments into Peru comparable to the tax sparing provision provided under the second sentence of article VIII § 4.

 

Consequently it is quite clear that royalties are not caught by article XII §2 and intentionally so. But, in addition hereto – and this indeed is the most important but also the most mysterious conclusion of the SAC – by virtue of its understanding and interpretation of the Spanish text of the treaty regarding the term “rédito de fuente”  the treaty includes an additional type of income which intentionally shall enjoy double non-taxation, namely capital gains!  But, instead of openly mentioning capital gains in the subordinate clause together with royalties to be protected from the subject to tax condition in article XVII § 2, the treaty negotiators have – one is tempted to say – ‘hidden’ this in the Spanish text! Or in other words, there are two limitations of the subject to tax provision in article XVII § 2, one explicitly mentioned in the subordinate clause regarding royalties and the other one in the Spanish text of the treaty regarding capital gains.

 

Apart from representing, generally, a quite awkward way of framing the scope of a treaty provision, a further comparison between the structure of the initialled draft of article XVII  § 2 of the treaty and the final text serves to discredit the idea of such a ‘hidden agenda’ in article XVII. Thus, in the initialled (English) text of the draft treaty not only royalty income under article VIII, but also dividend and interest income in articles VI and VII were excluded from the scope of the subject to tax condition. The article therefore read; “Subject to the provisions of Articles VI-VIII…” The reason herefor was that the draft treaty gave  Sweden the primary right of tax to all of these types of income and would credit any withholding tax due in Peru. Sweden was willing also in these cases to provide tax sparings. In the final agreement, however, which must have been the result of the exchange of letters that took place after the negotiation rounds, only royalties remained subject to this credit treatment whereas dividends and interest payments switched to the exemption method making them taxable only in the source state just like all other types of income (except royalties). This turnaround of the primary right of tax to dividends and interest income evidenced in the final texts of articles VI and VII, of course, had the consequence that the reference in the subordinate clause of article XVII § 2 to these articles must be deleted.  This again supports the validity of the explanation given above that any limitation of the scope of article XVII is embedded not in any  fancy interpretation of the Spanish term ”rédito de fuente en el Peru” but is explicitly spelled out and finally determined in the article itself, in this case by the introductory expression “subject to the provisions of article VIII…etc”. Therefore, if Sweden and Peru would  have had the common intention of excluding also capital gains from the scope of article XVII, they would have said so directly in the text, just as they did in the initialled draft regarding dividends and interest income.[17] Any other suggestion in this regard, with all respect, is based on an inadequate understanding of the structure of tax treaties.

 

Also one should  take into consideration that capital gains quite generally represents a type of income which, depending on the circumstances,  frequently is tax free according to domestic legislation and therefore most likely to be completely exempted from tax if the other state gives up its taxing rights in the treaty.  Therefore, if the negotiators had intended to allow double non-taxation with regard to capital gains it can be expected that this would have been very clearly indicated in the subject to tax article. ‘Surreptitiously’ tucking it away in an ambiguous Spanish text does simply not make sense.

 

A further point worth mentioning regarding the structure of the treaty is to take account also of its “systematic approach” and its “introduction and historical context” (see below). Adopting this recommendation it is especially interesting to note also what happened to article X, the capital gains article, during the course of the negotiations: In the initialled draft of that article the sole right of tax was assigned  (in conformity with the OECD Model Convention) to the state of residence of the seller of the capital assets which in this case would have been Sweden. But in the final treaty (see above) this was changed completely giving the right of tax only to the state where the assets were situated, which in this case meant Peru. At first glance this of course constituted a major sacrifice of Sweden’s domestic tax claims regarding capital gains, but only if the gains were taxed by Peru. If, on the other hand Peru did not tax the gain, which, at the time of the conclusion of the treaty, Peru did not, Sweden would retain its right to tax if article XVII § 2 covered such gains. But as Sweden - according to the SAC’s understanding of the Spanish text of this article - also excluded capital gains from the scope of article XVII  § 2 then there would be, not only a complete loss of Swedish taxing rights but it would also give rise to double non-taxation. Or in other words; Sweden during the course of the negotiations went from full Swedish tax on capital gains in the draft to full double non-taxation in the final agreement  derived by all Swedish resident tax payers!! In this context one should also remember that Sweden in the mid sixties was run by a social democratic government which at this time was in the process of abandoning a rather lenient tax regime regarding gains on shares allowing a five year tax exemption period and going to a system where gains would be taxed regardless of the length of the holding period. The suggestion that such a government simultaneously would make a tax treaty resulting in full double exemption is of such absurdity that it indeed must be considered “manifest”.
 

 

Reconciliation of the texts.

 

The above discussion has constituted an attempt to establish the meaning of article XVII § 2 of the Sweden-Peru treaty by applying the standard rules of interpretation in articles 31 and 32 of the Vienna Convention which under article 33 subsection 4 is the first measure to adopt  for the interpretation of a plurilingual treaty when confronted with a divergence between the different language versions. If such an investigation does not remove this difference of meaning article 33 stipulates that one should adopt “the meaning which best reconciles the texts, having regard to the object and purpose of the treaty”.

 

In this context it is interesting to observe that the SAC’s interpretation excluding capital gains from the scope of article XVII § 2 favours a restrictive understanding of that article.  Hereby the Court have – consciously or not – put their finger on an aspect of the ‘reconciliation process’ recommended under article 33 that has in fact been much debated in international treaty jurisprudence. Moreover, in this regard the SAC can also find support for its restrictive approach, and indeed from one of the most authoritative sources of international law in the world, namely the Permanent Court of International Justice in the Hague!

 

This emerged in the famous statement concerning interpretation of plurilingual instruments regarding the Mavrommatis Palestine Concession case, ((1924) P.C.I.J., ser.A, No.2) where the Court,  in connection with its examination of article II of the Mandate for Palestine conferred on His Britannic Majesty on July 24, 1922, declared:

 

“The Court is of opinion that, where two versions possessing equal authority exist, one of which appears to have a wider bearing than the other, it is bound to adopt the more limited interpretation which can be made to harmonize with both versions and which, as far as it goes, is doubtless in accordance with the common intention of the Parties”.

 

The debate on restrictive or extensive interpretation of plurilingual treaties in general and the Mavrommatis case in particular was given a lot of attention in the run-up to the Vienna conference. The International Law Commission (ILC)[18] in its commentary to the draft of the Vienna Convention rejected the idea that this case had intended to lay down a general rule that the more limited interpretation, which can be made to harmonize with both texts is the one which must always be adopted. Restrictive interpretation was justified in the special circumstances of the case, and it could therefore, in the opinion of the ILC, not be considered to call for a general rule establishing a presumption in favour of restrictive interpretation in case of an ambiguity in  plurilingual texts. But the Commission found that the case gave strong support for the principle of reconciliation of the texts. The aforementioned Swedish Foreign office report from the Vienna conference and the views of  Peter Germer in his article cited above also echo the opinion that there should be no general preference for a restrictive interpretation when interpreting plurilingual treaties.

 

The original version of the reconciliation  proposal presented by the ILC to the Vienna Conference did not refer to the object and purpose of the treaty but only mentioned that “a meaning which as far as possible reconciles the texts should be adopted”. This was however criticised by the United States during the course of the first session of the Conference suggesting that it be replaced by the formula “a meaning  shall be adopted which is most consonant with the object and purpose of the treaty”. The delegate of the United States explained that the proposal of the  ILC raised difficulty because it was merely an invitation to effect some sort of compromise but without any indication of the basis for their compromise. He also stated that in many cases a reconciliation of the texts was impossible.[19] After referring the problem to the Drafting Committee of the Conference and in what seems a compromise between the ILC proposal and the objections raised against it at the Conference, the final text of the Convention now provides that that when a  comparison of the authentic texts discloses a difference of meaning which the application of the standard rules of interpretation does not remove, the meaning which best reconciles the texts, having regard to the object and purpose of the treaty, shall be adopted. The original suggestion by the ILC that a reconciliation shall be effected thus remains as well as the US proposal that the object and purpose of the treaty should be the primary basis for such a reconciliation.

 

The reconciliation  process proposed by article 33, attractive as it may seem, is nevertheless problematic in practical application because it suggests that both the contracting parties should make some kind of sacrifice which normally would lead to an interpretation in favour of a (third) meaning  of the agreement never considered nor intended when the treaty was made. However, an examination of case law in these matters does not support such an approach. In the final analysis, see for instance the above cited Wemhoff, Bosch or Mavrommatis cases,  one will find that the courts and tribunals have always concluded that where a divergence exists between the authentic texts one or the other of them must prevail.  It is not possible to accommodate both the texts. And the same applies in this case.  One must make up one’s mind either that the term “income from a source in Peru” does in fact include a capital gain or that it does not. There is no third alternative.  However, by ‘amplifying’ the reconciliation process, as article 33 recommends in its final outline,  to take into account the object and purpose of the text, a useful mechanism is provided  to establish the intentions of the contracting parties.
 
In the analysis left above regarding the Sweden-Peru treaty it is suggested that the standard rules of treaty interpretation under articles 31 and 32 do in fact resolve the problems relating, in their context, to the understanding of the meaning of the term “income from a source in Peru”. If nevertheless the subsequent conciliatory process advocated in subsection 4 of article 33 of the Vienna Convention must be adopted and considering the intended object and purpose of article XVII § 2 in the Sweden-Peru treaty it seems obvious that this article of the treaty should be interpreted in such a manner that it includes also such capital gains that have not been subject to tax in Peru.

 

 

TREATIES WITH LESS DEVELOPED COUNTRIES AND TAX SPARING.

 

Also discussed during the court proceedings was a suggestion that one reason for giving article XVII § 2 a meaning whereby also capital gains would enjoy double non-taxation – this seems especially have been the focus of the dissenting members of the ARB, Wingren and Brydolf - was that the treaty was made (by Sweden) with an economically less developed country and thus reflected a wish to give Peru an economical favour in the nature of a matching credit or matching exempt provision (or tax sparing which is the term used by the OECD in its commentaries to the Model Treaty) regarding capital gains.

 

In this regard it should , however, be noted that tax sparing rules in our treaties, just like subject to tax rules, are also very detailed and carefully chiselled out, generally because they are of an irregular nature and represent exceptions from normal treaty practise (based on model conventions). Also, tax sparing rules are normally granted only with regard to interest or royalties and sometimes dividends paid from profits derived from specific and itemized activities[20]. Moreover, they usually apply only for a limited time period, of which article VIII § 4 of this treaty is a good example, and are subject to re-negotiation.  Sometimes, but very seldom, tax sparing may also be granted for taxes on permanent establishment income. Never, however, has  Sweden ever agreed to any  tax sparings with regard to capital gains on the sale of shares.[21]

 

Also, generally, the favours granted under tax sparing policies by Sweden will be considered only with regard to situations where the other state has adopted specific tax incentive measures for foreign investors for the promotion of its economic development but not to such cases where, according to a general domestic tax policy, a certain income or in this case a capital gain is not taxed. In the unlikely event that a tax sparing would in fact be granted in the future for capital gains taxes on stock holdings it will definitely be very narrowly construed and very clearly spelled out in the treaty text!

 

 

ARTICLE  3.2. (of the OECD model) AND TREATY PROVISIONS APPLICABLE BY ONLY ONE OF THE CONTRACTING STATES.

 

As mentioned earlier (and see also below) undefined terms of tax treaties based on the OECD Model may, under special conditions, be interpreted in line with the domestic law meaning of the state that applies the treaty. The main and natural justification herefor is of course that tax treaties are so closely related to the domestic tax systems of the contracting states. This specific character of tax treaties is something which sets them apart from other types of treaties, which mainly have been adopted exactly because there are no internal rules of the kind that these treaties are intended to address. Trade and peace agreements for instance are good examples hereof.

 

Another feature of tax treaties which to a certain extent also sets them apart from ‘ordinary’ international agreements is their reciprocal nature calling usually for equal tax sacrifices depending as the case may be on which state in the given situation is the state of residence of the tax payer and which state is the state of source of the income. This principle of reciprocity also means that the terms of the treaty shall be similarly interpreted by both the contracting states all of which of course  is in the interest of avoiding double taxation and preventing double non-taxation.

 

Sometimes, however, and not so seldom will this be specially mentioned in the Swedish government bills set forth for parliamentary approval, the rules of tax treaties are of such a ‘unilateral’ nature that it is quite clear that they can be applied by only one of the states or have been introduced solely in the interest of only one of them. This for instance is quite common regarding the avoidance of double tax provision. Indeed, when the contracting states adopt different methods for avoiding double tax, this is of course absolutely necessary. These types of provisions which have been adopted for ‘unilateral’ use by only one of the contracting states will often be closely geared to this state’s domestic rules of the same nature and/or , as mentioned, adopted to support a special requirement of that state. Therefore it seems quite appropriate that in these specific cases a preference be given to the internal law of that state when interpreting  the treaty provisions in question. This indeed is something, (see below under concluding remarks), which dovetails the statement made by the Swedish government when explaining their attitude to tax treaty interpretation in general and their approach to the language problems in particular.

 

Subparagraphs 1 and 2 of article XVII of the Sweden-Peru treaty are good examples of a treaty texts that have been drawn up in such a manner that they are to be applied separately by the two states. Subparagraph 1 is thus to be invoked  only by Peru when income has not been subject to tax in Sweden and subparagraph 2 applies in the reverse situation only to Sweden.

 

As shall be further explored in the following section the SAC in its application of the lex fori article has  concluded that the overriding purpose of treaty interpretation when faced with a term or an expression that has not been specifically defined in the treaty is to determine the common intention of the contracting parties and that the primary approach is to take into account the context in which the pertinent term appears. In this approach all circumstances shall be explored. It is suggested that the above mentioned ‘unilateral’ nature of a treaty provision shall be considered in this respect as an important element in establishing the context in which the relevant term appears. The result hereof will then be, which is also in line with the underlying purpose of article 3.2., that one shall allow for an interpretation in accordance with the domestic law meaning of the state applying the treaty and a meaning that complies with the purpose and intentions of this specific state when framing the provision in question. In this case regarding the Sweden-Peru treaty such an approach would of course also lead to an interpretation of the term “income from a source” that would include capital gains.

 

 

IN DUBIO, PRO TAX PAYER OR PRO FISCUS?

 

In his dissertation on the taxation of  cross-border partnerships (Volume 9, Doctoral Series. International Bureau of Fiscal Documentation – Academic Council) Jesper Barenfeld brings to attention a quite novel idea regarding tax treaty interpretation, the main conclusion beeing that (page 35) “in cases where neither the treaty text nor any of the supplementary sources recognised under the Vienna Convention on The Law of Treaties give a clear answer, much speaks for interpreting a treaty according to the alternative most favourable to the tax payer. Such an approach supports the objective of protecting the individual’s rights against the state’s powers to levy taxes in unclear cases. It also advocates the important aim of avoiding international juridical double taxation”.

 

An interpretation along these lines is of course desirable because it, too, supports an approach in accordance with the object and purpose of the treaty. However, there are several reasons why, nevertheless, it should be applied with great caution. Firstly it is not recognised by the Vienna Convention, (although, as just mentioned an observance of the object and purpose criteria of articles 31 and 33 when interpreting a tax treaty will in itself strengthen the taxpayers situation). Secondly it doesn’t answer the question which of the contracting states that should make the sacrifice. The state of residence of the tax payer or the source of the income state.  A third and most important reason for rejecting a pro tax payer approach as a general principle in tax treaty interpretation is that these treaties, (also in contrast to most other types of  international agreements), provide a special provision enabling the taxpayer, if a court decision is disadvantageous to his interests and results from actions that are not in accordance with the treaty, to apply for a competent authority ruling of his case under the mutual agreement procedure of the treaty (article 25 of the OECD Model).

 

Consequently, and in contrast to the suggestion by Mr. Barenfeld, the court, when in doubt, should adopt the opposite approach in favour of ‘in dubio pro fiscus’. This is also in accordance with the conclusion made by the Swedish government in the Nordic tax treaty bill (1989/90:33, page 33) mentioned above , that if  Sweden may impose taxation under its domestic laws and has not explicitly given up this tax claim in a treaty, then it shall be taxed in Sweden. If this leads to double taxation it should be removed by competent authority resolution according to the mutual agreement procedure.

 

Also worth considering is that a tax treaty, despite its formal incorporation in domestic Swedish law according to our monistic principle of ratification of international agreements, remains an agreement between the two contracting states which does not confer upon the taxpayer any categorical or absolute entitlement in all cases to the benefits of the treaty. In fact, the very presence of the mutual agreement article actually reflects the understanding that the court procedure will not always be adequate for a resolution of the dispute and that this is the reason for giving the taxpayer a further instrument for a solution of his problem. Again, however, it must be remembered that not even the competent authorities are obliged to reach a solution that does away with the double taxation. Mr. Barenfeld also seeks support in the famous maxim originating in Roman (criminal) law “in dubio mitius” and refers also to the position taken by the European Commission on Human Rights in the Golder case determining that the interests of the individual should override the sovereignty of the State(s). It should be remembered, however, that this case addressed the question of the human rights of Mr. Golder whereas, in contrast, the protection from double taxation in a tax treaty, as much as it is desirable, does not rank as such a human right.

 

Mr. Barenfeld also points out, and quite appropriately, that his ‘pro taxpayer principle’ does not apply in situations where an interpretation would lead to a double exemption “as this is not a benefit that the tax payer has any reasons to expect” (page 34). In such cases it therefore seems especially well-suited to apply a pro fiscus principle considering also that a  pro taxpayer decision by the court allowing a double non-taxation naturally would never be brought by any (sober) tax payer to a competent authority procedure.

 

 

CASE RÅ 1987 ref. 162

 

This  case which has been referred to several times previously in this article is a landmark decision in tax treaty interpretation matters in Sweden regarding the (terminated) 1968 Sweden-UK treaty . It includes a statement that can be regarded as the ‘bible’ of tax treaty interpretation addressing the much debated understanding of the lex fori article, (number 3.2 of the OECD Model Treaty) which has been referred to above.  The text of this article in the  Sweden-Peru treaty reads (Article II § 2): “In the application of the provisions of this Agreement by one of the Contracting States any term or expression not otherwise defined shall, unless the context otherwise requires, have the meaning which it has under the laws of that  State relating to the taxes which are the subject of this Agreement”.

 

This is how the SAC interpreted the corresponding treaty provision in the 1987 case, (the full text of which was  included in the file of this case for consideration by both the ARB and the SAC):

“Domestic tax rules shall be applied only unless the context otherwise requires. The meaning of this restriction….may be summed up as follows: Where a term of a treaty, as used in any treaty provision, does not give a  clear indication of its meaning it is necessary to try to establish the intentions of the contracting parties. In so doing guidance should be sought from the terminology of the treaty as a whole, its structure and systematic approach, the function of the article in question , its introduction and historical context  as well as other relevant circumstances. Only if such an investigation does not lead to any result shall recourse be had to the meaning of the term under domestic law of the state applying the treaty.”

This very eloquent proclamation has been repeated verbatim in many subsequent cases, for instance RÅ 1996 ref. 84. Thus, as far as is known, Sweden is the only country in the world, whose Supreme Tax Court  has given a separate and specific interpretation of this rule of interpretation embedded in each and every tax treaty! The general message in this ruling is that, in the quest for the intentions of the parties, the meaning of the term in its context is the supreme determining factor of the interpretation process. And all the different ingredients that shall be considered in the determination of the context are listed in the ruling! Thus, only as a very last resort measure, should the internal law meaning be adopted. The very broad picture that is painted of all the ‘colours’ that should be used by the interpreter goes so far that it should always be possible to determine the context and that, consequently, it should never be necessary to revert to any internal law meaning. [22]  Important to observe  is of course that the Vienna Convention itself also stresses (in article 31) the importance of the context in which the term appears in the treaty and also that it leaves no references to any internal law meanings of the contracting states.

 

As noted above the undefined term of the Sweden-UK treaty that was under scrutiny in the 1987 UK was also “income from a source” as it was to be applied to a capital gain on the alienation of shares. Exactly the same problem as in this case! Another striking similarity to the present case was that it also occurred in the context of an anti double non-taxation provision namely the subject to remittance rule under which Sweden agreed to treaty relief (as source state) of such gains only on the condition that the “income” had been remitted to the UK. The underlying purpose of this rule was of course equally clear. In its ruling the SAC (with a majority of 3-2) came to the conclusion, however,  that the disputed term should have the meaning it had not under Swedish law but indeed under UK domestic law excluding capital gains from the scope of the article. In the present case the justification for the outcome reached by the SAC is not  the Peruvian law meaning but rather the Spanish language version. But the main and quite ironic similarity between the two cases is that Sweden, as suggested by the SAC, by using the rather innocent and common tax treaty term “income from sources”  has deliberately excluded capital gains from taxation, in the 1987 ruling from the application of the subject to remittance clause and in this ruling  from the subject to tax clause, in both cases giving  rise to an  intentional  double non-taxation in articles of the treaties that were specifically designed to prevent such double non-taxation! A further similarity of the two cases is that the SAC in both of them reversed the rulings of the ARB, which both led to the prevention of double non-taxation.

 

In what was their primary argument, and which indeed is relevant for the discussion of this case, the majority of the SAC in 1987, in support of their ruling, referred to the fact that the treaty negotiations had been conducted in English also of course leading to an initialled draft in that language. Such an approach, giving preference to the language in which the treaty was negotiated, however, as mentioned above, is strongly and categorically rejected by all authorities on the interpretation of plurilingual treaties. Peter Germer, for instance, concluded that such a principle, according to the Vienna Convention, cannot be considered a valid rule. Once it is established that the different language versions are authentic, the interpreter is bound to give effect  to the principle of the equality of the texts, and it would be a clear violation of that principle to consider the original language version to be superior to the final authentic texts. However, concludes Germer, “an examination of the preparatory works, including an original draft language version of the text may be considered, according to article 32 of the Vienna Convention, to display the causes of a divergence between the different authentic language versions and thus serve to establish the meaning intended by the parties to be attached to the provision in question”

 

In this case the initialled English draft of the Sweden-Peru treaty serves excellently as an instrument to display the cause of the divergence and for establishing the meaning intended by the parties to be attached to the meaning of  the term “income from a source”.[23]

 

Most importantly,  the 1987 ruling made it abundantly and “conclusively” clear that the term “income from a source” if the Swedish domestic law meaning or the Swedish text of the treaty had prevailed,  would have applied also to capital gains resulting from a sale of shares.

 

But despite the similarities between the two cases and  the fact that the 1987 ruling - all 14 pages thereof -  was (physically) included in the case dossier and presented both to the ARB and the SAC for consideration, there is no reference whatsoever  thereto by any of theses courts nor by any of the litigating parties! (The ARB and the litigants on the other hand may be ‘excused’ herefor as the relevance of the meaning of “income from a source” did not emerge before the SAC spotted the language problem in the two texts.)

 

 

 

 

CONCLUDING REMARKS:

 

The above analysis of the Sweden-Peru case does not leave much room for praise of the SAC ruling. It is especially disappointing to see how casually the Court both in 1987 and again in this case have ignored the object and purpose of the treaty provisions they have set out to interpret, brusquely brushing aside the double non-taxation issue. Indeed, the very fact that the Court have openly mentioned that it is immaterial for their conclusion that Peru did not tax the gain – a statement, moreover, that was quite unnecessary for justifying their conclusion -  illustrates very clearly their low ranking of the object and purpose of treaties in the interpretation process. This is so much more surprising considering the total ‘devotion’ in which the SAC, generally, holds the Vienna Convention.

 

One reason for the Court’s disregard of the Vienna Convention’s strong respect for the object and purpose in treaty interpretation may be that it calls for an adherence to the teleological method of interpretation [24], and this method has a very low standing with regard to interpretation of Swedish domestic law in general.[25]

 

But a tax treaty is different from domestic tax law despite the fact that it has been incorporated therein! It remains a contract between two (or more) parties and in contract law the intention of the parties is a most important ingredient calling for a teleological approach. Also, whereas in a normal contract there may be conflicting interests and objectives obscuring the intentions of the parties, this is not so pronounced in a tax treaty (contract) with its reciprocal nature always striving to promote cross-border business and to prevent double taxation which, in the final analysis, is to the benefit of all parties.

 

A general reason justifying a different approach to the interpretation of tax treaties in comparison to interpretation of domestic tax laws follows also from the different structure and scope of these texts. Whereas, on the one hand, domestic tax laws may comprise dozens, yes, sometimes hundreds or more pages of detailed provisions, tax treaties on the other hand are only some 40 to 50 pages long. Also, they  must adapt to changes of the domestic tax laws and consider transactions and investment strategies which did not exist at the conclusion of the treaty. This, indeed, is a reason why tax treaties sometimes wittingly may be left ambiguous or inconclusive permitting a  more flexible approach by the judge.

 

This special nature of treaty texts as compared to domestic laws and, indeed, also the problems resulting from the plurilinguility of the treaties is reflected too in the above mentioned government bill regarding the Nordic Tax Treaty (page 93) as well as certain other treaties e.g. the treaties with Barbados and Namibia. This is what the finance minister said in those government bills : “Unintended tax benefits may also arise because the provisions of  tax treaties often lack the exactness of similar rules in domestic law. One of the reasons herefor is that the tax systems of the contracting states display a number of differences which places a heavy burden on the drafting of the specific treaty provisions. This may have the result that these provisions, from the point of view of one of the contracting states, cannot always provide an optimal picture of their underlying intentions. A further difficulty follows from the fact that most treaty negotiations...…… are conducted in a language that is alien to the contracting states and that the treaties are also authenticated in such a language. This can have the effect that words and expressions which in Swedish or the foreign language have an exact meaning must be translated and that therefore the exactness may be lost. This can also occur where, due to a  unilateral Swedish requirement, a specific provision referring to Swedish domestic law is adopted in the treaty. Even if all tax treaties concluded by Sweden are drawn up in both one or more foreign languages together with Swedish or at least are translated into Swedish this may have the result that the Swedish text cannot reflect the exactness of the Swedish domestic law meaning.”

 

This implies that it is simply not possible always to express one’s exact or complete intentions in a treaty that has been negotiated in a foreign language and/or authenticated in two or more languages. Therefore, when considering the “ordinary” meaning of the terms of a treaty, one must allow for a more ‘imaginative’ interpretative method than the uncompromising literal approach applied to domestic texts. This, it is suggested, is what is embodied in the requirement that a treaty must be interpreted “in its context” and in “the light of its object and purpose”, breaking loose from the straightjacket of a rigid linguistic interpretation, embracing instead the eloquence of the Advocate General Maurice Lagrange’s  quest for “l’esprit du texte”!

 

But, alas, both the 1987 case and now again the outcome of the interpretation of the Peru treaty show very little of  all this.  And the current is running  hard against the attempts of our treaty negotiators to prevent  double non-taxation! This is quite lamentable as double non-taxation in tax conventions is especially serious, undermining  the integrity of the tax system on the whole and tax treaties in particular. Double non-taxation is justly considered unfair by the ordinary tax payer who grudgingly but loyally shoulders his fair share of a single tax burden. But this loyalty will evaporate fast when he realises that international agreements concluded for the specific purpose of preventing of double taxation does away with taxes altogether for some tax payers! When, in addition, he learns that such double non-taxation is concocted, as suggested by the SAC in this case, by deliberate actions taken by his legislator/negotiator, he may well stand up in anger, or, as someone suggested, call for the appointment of new SAC judges!

 

A further reason for disappointment with the SAC ruling is the precipitate and perfunctory nature of the Court’s administrative management of the case which also seems to underlie its remissness of the interpretation principles set forth by the Court itself in 1987 to “seek guidance from the terminology of the treaty as a whole, its structure and systematic approach, the function of the articles in question, its introduction and historical context as well as other relevant circumstances”. Thus, the late discovery of the perceived divergence between the Swedish and the Spanish texts left no opportunity for an investigation of the English text in the initialled draft of the treaty, the linguistic analysis of the Spanish text is, as it appears, cursory, the importance of the structure and ‘design’ of subject to tax rules has been left unexplored and vital information in the OECD Model Treaty has been ignored. And, most disturbingly, one is left with the impression that article 33 of the Vienna Convention’s rules of interpretation of treaties authenticated in two or more languages has been completely over-looked.

 

The outcome of the 1987 decision, prompting hordes of Swedish shareholders to move to the UK to take advantage of the double non-taxation situation following therefrom, gave rise to Swedish tax base losses of immense proportions (before the leak was finally stopped by the conclusion of a new treaty). Since January 2004, and as mentioned earlier, Swedish corporate investors, under domestic tax law, now enjoy a participation exemption regime whereby gains on controlled shareholdings domestic or foreign are generally exempt from tax. This ruling, therefore, is no longer of any importance to this category of tax payers. Swedish investors on the Lima stock exchange can, however, reap benefits from the ruling. Also, and more importantly, as the ruling and its double non-taxation effects, together with the additional facts [26]that Peru is on the “white-list” for controlled foreign corporation taxation purposes, applies also to individuals resident in Sweden, it has let loose a wave of tax planning which could lead to serious abuse of the treaty. The extent hereof  has in fact reached such proportions that the Government  has now given notice of termination of the Peruvian treaty which will become effective 1 January 2007. [27] This is the second time in a short period that the legislator has been compelled to counteract the consequences of the SAC’s  interpretation of our tax treaties. Recently the Swedish credit of foreign tax law had to be changed in order to ward off the disastrous effects of the SAC’s ruling in the so called Alecta case, RÅ 2001 ref. 46, denying foreign tax credit to Swedish partners of  tax transparent partnerships.

 

The double non-taxation loop-hole lurking in the Spanish text of the Peruvian treaty  - it is safe to suggest -  would never have been detected without the ingenious ‘assistance’ of the SAC. Nor would any tax practitioner, however frivolous,  have dared to exploit it without a binding ruling . Or, as one observer exclaimed: “The Supreme Administrative Tax Court of Sweden -  the international tax planner’s best friend!”

 

Stockholm, May 2006

 

peter@sundgren.net

 

 

 

 

 

 

 


[1] See also by this author British tax review 1990 No.9, “Interpretation of tax treaties – a case study.”
[2] Under participation exemption legislation now in force the sale of  shares, both Swedish and foreign, held by controlling Swedish companies are exempted from Swedish tax. Swedish resident individuals,  however, remain subject to capital gains tax on the alienation of shares.
[3] However, according to legislation introduced in 2003, Peru imposes a 30% tax on capital gains derived by foreign resident corporations or other legal persons if the gain applies to stocks or other bearer securities issued by companies, security investment funds or trust funds created or established in Peru unless the gain accrues before 1 January 2007 and the sale is made over the Peruvian stock market. Foreign resident individuals are also subject to the capital gains tax for alienations of Peruvian shares occurring as of 1 January 2007. The issuing company etc. is liable for the payment of tax of the non resident owner. All Peruvian residents pay the 30% capital gains tax as of 2007.
[4] Mr Ersson is also an erstwhile and longtime tax treaty negotiator of Sweden.
[5] This statement is of some interest because it rejects the declaration put forward by the Swedish Government in many of its bills regarding tax treaties, see for instance pages 93-94 of Riksskatteverkets presentation (2000) of the Nordic Tax Treaty, according to which the Vienna Convention’s interpretation rules “aim at resolving disputes regarding the meaning of the treaty arising between the contracting states in order to determine their, i.e. the states’ intentions on specific treaty provisions and not – as far as tax treaties are concerned – to govern the relationship between the tax payer and the State”.
[6] ”The Interpretation of Tax Treaties with particular reference to Article 3 (2) of the OECD Model by John F Avery Jones (United Kingdom), Charles J Berg (Australia), Henri-Robert Depret (Belgium), Maarten J Ellis (Netherlands), Pierre Fontaneau (France), Raoul Lenz (Switzerland), Toshio Miyatake (Japan), Sidney I Roberts (United States), Claes Sandels (Sweden), Jakob Strobl (Germany), and David Ward (Canada).
[7] See for instance Claes Sandels/Harry Margulies, Skattenytt 1984 page 412 ff.
[8] Notable in this context is the fact that the Swedish tax treaty model is made in English

[9]Finansdepartementets Konseljakt den 27 Januari 1967, nr 35.

[10] And it remained Peru’s only bilateral treaty until 2001 when such a treaty (not yet in force) was concluded with Chile. Effective January 1 2004 Peru now also has a bilateral tax treaty with Canada.
[11] Sometimes, subject to tax rules may function also to the benefit of the source state, which may be allowed to tax an income that is not subject to tax in the state of residence of the taxpayer. See for instance article 26 subparagraph 5 of the multilateral Nordic tax treaty.
[12] The background hereto is as follows: Finland provides certain pension benefits that are tax free (which of course is reflected in the level of payments of these benefits). Under the subject to tax rule in article 25 the other contracting states may tax all income that is untaxed by Finland if the recipient is resident in any of the first mentioned states. For a recipient of such pensions resident for instance in Sweden this taxation would of course be quite unfair compared to beneficiaries resident in Finland receiving the same pension free of tax. Therefore article 25 subparagraph 3 provides that the subject to tax rule just mentioned shall not apply with regard to these specific pension payments.
[13] It is noteworthy that Sweden, even when relieving double tax by way of the credit method, has subject to tax rules! An example hereof is the treaty with South Africa (article 26 subparagraph 2), which, at the time of conclusion of the treaty, also had a territorial tax system. The purpose of the subject to tax rule in this case is to catch non-taxed third state sourced income in cases where the tax payer has double residence under the treaty.
[14] An uncertainty may arise in this context when the income has been indeed effectively taxed but where due to a deduction of expenses or losses no tax is levied.
[15] The so called 6-month rule under Swedish domestic tax law is an example where the  tax exemption in Sweden rests on the condition of an effective tax assessment treatment in the other country.
[16] The main interpretation problem regarding the distributive rules is the classification of the nature of the income. For this reason these articles always provide comprehensive definitions in this respect.
[17] An additional very clear example of this exactness of the ‘construction’ of a subject to tax clause, leaving no uncertainty about its scope, can be studied in article 26 of the multinational tax treaty between the Nordic countries.
[18] The ILC was the (very influential) secretariat appointed by the United Nations to draft the Vienna Convention.
[19] Doc.A/CONF.39/C.1/SR.34, at 10 (April 23,1968).
[20] Article 22 subparagraph 5 of the Sweden-Mexico treaty is a good example hereof.
[21] Usually, this situation will not occur in tax treaties because only very seldom  do source states under their domestic tax regimes impose capital gains taxation on foreign residents concerning movable property and most treaties assign primary and sole taxing rights to the state of residence of the tax payer.
[22] The lex fori rule was at the focus of attention at the 1993 IFA Congress in Florence. I was on the discussion panel of this subject explaining the 1987 case to the delegates. I gave expression to my deep satisfaction with the Court ruling, relegating  the application of the internal law meaning of undefined treaty terms to very last resort use. But I went one step further recommending a complete rejection of the use of internal law meanings altogether because a) it had no place in the Vienna Convention’s rules of interpretation and b) mainly because, considering that the internal law would only be used when there was a difference between the internal meanings of the two contracting states, it would automatically and always give rise to either double taxation or double non-taxation! If, on the other hand, the internal law meanings in the two states was the same, I argued, there would logically be complete and automatic harmony of the meaning of the term also for the purposes of the treaty. In such a case it would of course be unthinkable to adopt a different meaning. The two domestic law meanings would in fact be considered the ordinary meaning consistent with the Vienna Convention’s main rule. Thus, I concluded, article 3.2 was not only unnecessary but also outright detrimental to treaty interpretation in that it, as aforementioned, always leads to double taxation or double non-taxation. In my enthusiasm over this conclusion I even suggested that the Congress vote for the abolishment altogether of article 3.2! The ballot, however, resulted in an overwhelming and dismal defeat of this proposal.

 

[23] It is of course equally appropriate, if a treaty is drawn up in numerous languages, as for instance the Nordic treaty, to try to settle the cause of the problem by comparing all the different texts. If for example, in such a comparison, a divergence of language appears in only one text, there would be a logical inclination to accept the understanding collectively reached by the other states.
[24] In a footnote to Series B NR 10, 15 f, Serie A/B Nr 68, 60 the International Court in the Hague, with regard to the interpretation of international texts has also expressly declared: “Es handelt sich um eine ‘teleologische Auslegung’.
[25] The SAC’s  practise in this regard is, however, not consistent. Thus in RÅ 2004 ref. 29, and basing its interpretation on a very  strong purposive element, the Court, (overruling the decision of the ARB), came to the  conclusion, in what incidentally is a case on international taxation, that a foreign (Norwegian) foundation was a ”foreign juridical person” under the very carefully  formulated rules of chapter 6 paragraph 8 of the Income Tax Code (ITC)  meeting the criteria of being an ”association”  because it had been ”formed under the rules applicable to associations”.   The Court in this context underlined the necessity of observing the rules (in chapter 6 paragraph 7 of the ITC) of tax liability of non-residents concluding, that if a foreign foundation did not constitute a ”foreign juridical person” it would not be liable to tax in Sweden even if e.g. it derived business income from a Swedish permanent establishment. Nor could taxation normally be imposed on its partners, which should be observed when determining the requirement under subparagraph 3 of chapter 6 paragraph 7 ITC (stipulating that the partners should not freely be able to take possession of the property of the association). This argument was supported by the fact that;  “The explanation to the structure of the definition…. occurred in the context of a legislation the purpose of which was to determine whether the foreign association or its partners should be liable to taxation in Sweden. With this in mind it cannot be understood that the absence of partners in a foundation or a non-profit organization (ideell förening) excludes its character as a foreign juridical person. Instead, the provision focuses on the situation where there are partners ”. Or, in plain language: as it could not have been the purpose of the legislator that foreign foundations nor their  beneficiaries should escape tax liability on , for instance, Swedish permanent establishment income, the law should be interpreted in such a way that the purpose of the legislation is fulfilled. Thus, seemingly, the definition in chapter 6 paragraph 8 of the ITC of the term “foreign juridical person” applies only to such foreign ‘corporate’ persons that have partners (delägare). With regard on the other hand to persons such as foreign foundations, trusts, non-profit organizations, death estates (and possibly a few others), if they are formed under rules applicable to associations, that do not have partners the Court has detemined that they too shall be regarded as “foreign juridical persons” (at least) for the purpose of preventing that they should otherwise escape Swedish tax liability of non-residents.  It must be assumed that the legislation referred to by the SAC regarding the taxation of foreign associations or its partners is our 1991 controlled foreign corporation (CFC) taxation regime. This leads to the conclusion that Swedish resident beneficiaries of low or no taxed foreign foundations and trusts are not subject to CFC taxation but that the foundation/trust, as just mentioned, is liable to tax for Swedish source income. (Considering, however, that foreign trusts, as the case may be, are not “persons” at all it cannot be ruled out that Swedish resident beneficiaries or  settlors are considered the ‘beneficial owners’ of the trust property and thus liable to tax on its income and capital.)
[26] Negotiations for a new treaty have started but the Government is not prepared to let the present situation prevail for the number of years it will take to conclude such a new treaty. (The government also concluded that the treaty in practice is useless to Swedish investors as it does not provide for any reduced source taxes.)

[27] For an analysis hereof see, by this author (in Swedish), IUR-Information (IUR-Institutet för Utländsk Rätt) no. 11/2001, “Förhandsbesked orsakar dubbelbeskattning för miljonbelopp”.