I would like to thank Ankit for inviting me to write a few guest posts for this blog. In this post, I wish to look at some aspects of non-discrimination protection in international tax law.
1. Introduction
Non-discrimination provisions are perhaps among the lesser-used provisions in tax treaties. Under the OECD Model Draft, the principle of non-discrimination is covered under Article 24. A survey of the reported cases will reveal, however, that the provision is hardly ever used. Two questions prominently arise in this regard. First, when will it be said that a provision of law differentiates between two groups in a relevant sense? And secondly, when would that differentiation amount to discrimination?
The first issue essentially turns on the identification of a comparator – treatment against the assessee must be compared vis-à-vis treatment against ‘X’ entity. That ‘X’ entity is the comparator. How is this comparator to be identified? This issue is addressed in a judgment of the Pune Bench of the ITAT,
Daimler-Chrysler, about which I have written a two-part post available
here and
here.
The second question turns on the standards applied in assessing the content of the protection against discrimination. I will concentrate on this second issue, but before that, a slight detour to the rationale of non-discrimination provisions may be relevant.
2. The Broad Object and Rationale
Non-discrimination provisions in tax treaties are in general intended to eliminate tax discrimination in certain precise circumstances. The provisions seek to “balance the need to prevent unjustified discrimination with the need to take account of legitimate distinctions” based on, for example, difference in liability to tax or ability to pay [OECD, Model Tax Convention on Income and on Capital (Commentary) (2008)].
A slight historical tour may be useful in further contextualizing non-discrimination protections in tax treaties. Non-discrimination provisions first made a concrete entrance in the sphere of international economic law in several Treaties of Friendship, Navigation and Commerce. These treaties are the precursors to modern bilateral investment treaties; and were intended to promote, encourage and protect international trade and commerce, and foreign investment. It was from here that the League of Nations borrowed the non-discrimination concept into tax law. The Fiscal Committee of the League, in its Mexico Draft in 1943 and in its London Draft in 1946 specifically included a clause against discriminatory tax treatment of foreigners. [See generally: Kees van Raad, Non-Discrimination in International Tax Law (1986)]
The stated objective of this provision at this juncture was to prevent discriminatory treatment of tax payers having their fiscal domicile in another country and in order to aid the protection of foreign capital in a domestic country [In Re P. No. 6 of 1995, (1998) 234 ITR 371 (AAR)]. This development was reiterated by the OECD in 1958; and then in 1977 [K. van Raad, “Non-Discrimination” (1981) British Tax Review 43].
A separate rationale was subsequently enunciated in the OECD Model in so far as ownership non-discrimination is concerned. This rationale is not in substitution of, but is in addition, to the preceding one. This ‘new’ basis is to ensure equal treatment for tax payers residing in the same State, no matter where the capital supporting the tax payer is coming from. In sum, the ownership provision (as per the OECD Thin Cap study, 1987) “aims broadly at preventing tax protectionism – i.e. the deterrence by tax measures of investment from outside the country.” The principle behind this is that a State ought not to tax local enterprises owned by residents of another State more harshly than it taxes local enterprises owned by residents of the same State. This has in general been “a widely accepted feature of international economic agreement for decades.” [See: M. Bennett, “Non-Discrimination in International Tax Law: A Concept in Search of a Principle” (2006) 59 Tax Law Review 439]. The rational is thus to protect foreign owned capital. This broad rationale of the non-discrimination provision will be useful in resolving more concrete issues, to which this paper will now turn. [See generally: J. Avery-Jones, “The Non-Discrimination Article in Tax Treaties” (1991) British Tax Review 421].
Importantly, a non-discrimination article in the 1977 OECD Model was based purely on this preceding history and rationale. Subsequent OECD Models do not change this basic rationale. Furthermore, in relation to the non-discrimination article, the UN Model specifically referred to and based itself on the 1977 OECD Model. Thus, modern-day non-discrimination clauses can all be traced to broadly the same rationale. Of course, the travaux in individual treaties may suggest otherwise; but broadly, it would be safe to assume a general rationale for all non-discrimination protections in tax treaties.
With this broad objective of the ND clause in contemplation, I move back to the issue of when the non-discrimination articles are said to be breached – what exactly is this discrimination?
3. When the ‘differentiation’ amount to ‘discrimination’?
Commentaries and commentators are practically unanimous on one point – that all differentiation does not ipso facto result in discrimination. Judicial decisions also echo the same thoughts. In one view, “discrimination means distinguishing between persons adversely on grounds that are unreasonable, irrelevant or arbitrary.” [Arnold and McIntyre, International Tax Primer (2002) 128]. However, whether a particular ground is unreasonable, irrelevant or arbitrary is simply “a matter of judgment” in the sense that it finally depends on what legal test from several available the judiciary chooses to apply [See van Raad, cited above].
Unfortunately, the question of what exactly is the standard to be applied has, however, not been considered in great detail by Courts and tribunals. One of the few cases which stands out in this regard is Automated Securities v. ITO, ITA No. 1758/PN/2004. In this case, the Tribunal adopted a test based on the constitutional right to equality – Article 14 of the Constitution of India. Indeed, the Tribunal specifically cited leading constitutional law cases on the point [Kedarnath v. State of West Bengal, AIR 1953 SC 404; State of West Bengal v. Anwar Ali Sarkar, AIR 1952 SC 75]. The other decisions which make a non-discrimination analysis can also be explained on the basis of this test; and no decision specifically leys down a different test. [Reference may be made, in particular, to the following: Chohung Bank v. DCIT, Mumbai Bench, Income Tax Appellate Tribunal; Herbalife v. ACIT, (2006) 103 TTJ (Del) 78; Metchem Canada v. DCIT, (2006) 284 ITR 196 (Mum). Also see: ABN Amro Bank v. JCIT, (2005) 96 TTJ (Kol) 1041]. Consequently, the Indian cases on the point seem to suggest the application of an Article 14 test to the issue.
Thus, one can see from the perspective of the (rather limited) Indian practice of international taxation which considers the issue in detail, that the Article 14 test is applied.
It must be noted at this juncture that it is extremely difficult to sustain an Article 14 challenge to economic and taxation laws [Southern Technologies v. JCIT, Civil Appeal No. 1337/2003, decision dated January 11, 2010]. One issue which needs to be considered is whether this adoption of the Article 14 standard by Indian decisions in interpreting the non-discrimination article in tax treaties is in harmony with the principles of treaty interpretation.
To begin interpreting the non-discrimination article for present purposes, one can briefly outline the principles of interpretation of treaties. The customary international law principles dealing with the interpretation of treaties find themselves codified in the Vienna Convention on the Law of Treaties. Article 31 requires a treaty to be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose. It is implied in this that the principle of effet utile (which suggests that every provision of a treaty must be given some meaning and not be rendered meaningless) applies. Furthermore, the general rule has been extended to imply that interpretation in harmony with surrounding customary international law is to be preferred. Article 32 deals with the supplementary rules of interpretation; used when the meaning under Article 31 is unclear, or used in order to confirm the meaning arrived at under Article 32. [I have avoided heavy citations – but these propositions may be verified by reference to any standard international law textbook].
There is nothing whatsoever in these principles to suggest that the Article 14 test is apposite in the context of interpreting the non-discrimination provision in treaties. Indeed, using an Article 14 test perhaps violates the principle of effet utile – if the test is the same as an Article 14 test, then a separate provision would not be required at all. Both resident and non-resident assesses, both legal and natural persons, are entitled to challenge the vires of a law as violating Article 14. If all assessees in any case have this protection as a matter of constitutional law, why would India and its treaty partners have inserted a non-discrimination article at all into the DTAAs? Furthermore, what is the basis which suggests that the parties to a tax treaty would have intended the domestic constitutional meaning of one party to prevail? Such an assumption does not seem appropriate to begin with, unless the travaux of the particular treaty clearly so indicate. The better approach may be to begin from the principles of treaty interpretation, without being coloured by the fact that non-discrimination provisions exist in Article 14 of the Constitution.
On a purely literal meaning, arguably, discrimination is not easily distinguishable from differentiation. Of course, in law the two are different concepts – but that difference is not easily traceable to the literal meaning of the word. A standard dictionary, for instance, defines ‘discrimination’ as ‘the act of making or recognizing differences and distinctions’ and as ‘a making a difference in favour of or against’ [see van Raad, page 7, citing to several dictionaries]. The verb form ‘discriminate’ is defined as ‘to show or make a difference in treatment’ [For instance, see: World Book Dictionary (1989) 600.].
The legal difference between ‘discriminate’ and ‘differentiate’ arises more from the context in which the non-discrimination protection exists; rather than from the inherent meaning of the words themselves. Thus, in order to appreciate what exactly the difference between the two concepts is, one must refer to the specific context in which the protection exists. Hence, merely borrowing a test from a different area of law may not be appropriate – one must seek to look to that area which most closely approximates the object and purpose of non-discrimination protections in tax treaties. In this view, looking at the literal meaning of the word in light of the object and purpose of non-discrimination protections, the Article 14 test appears to be misleading.
Where else does one look, then? I will try to provide an answer in a subsequent post – meanwhile suggestions from other readers and commentators are most welcome.