I want to say a bit more about Nicolas Diebold's SIEL conference paper "Non-Discrimination and the Pillars of International Economic Law – Comparative Analysis and Building Coherency". I've read it more closely now, and I like it even better than when I first skimmed through it. Let me go through some of the key points he makes.
First, he argues that the lack of clarity in non-discrimination standards is a serious problem:
the current situation entails a number of shortcomings. Most importantly, contracting parties have no possibility to anticipate the consequences when entering into an international economic treaty. By the same token, private individuals – to the extent the treaty empowers them to assert their rights under the agreement – are virtually left in the dark when assessing their rights and risks prior to an investment decision. The scope, substance and standard of a non-discrimination obligation – and nota bene many other obligations – depends on the interpretation of the individual arbitrator who happens to be appointed to rule on the specific dispute. Moreover, even once the arbitral tribunal made its ruling by applying a specific treaty, the concerned parties have no guarantee that a second arbitral tribunal will follow the same interpretation. Consequently, the parties to a dispute – whether in state-to-state or investor-state arbitration – have no or very little guidance on how to make their claim, how to present their legal arguments and, most importantly, what evidence will be relevant.
I absolutely agree. Non-discrimination is a core legal standard in international economic law. We can never have complete certainty about what the law is (there are competing views, and ideas change over time), but we should know a lot more than we do currently. (As noted here, there are some cases coming up that could clarify things in the WTO context).
Next up, when describing the "fragmented standards and terminology" for "less favourable treatment," he discusses what is often referred to as the "best treatment" approach:
The non-discrimination principle may also be interpreted as an obligation to grant the best treatment accorded to any domestic market participant. Following this approach the non-discrimination obligation is already breached if one individual foreign market participant receives treatment that is less favourable in comparison to any individual domestic market participant (NT) or to any foreign market participant from different origin (MFN). For instance, less favourable treatment occurs if a measure negatively affects only 1 out of 100 foreign market actors, even if 99 out of 100 domestic market actors are also negatively affected. Consequently, the non-discrimination principle becomes an obligation to treat all foreign market participants equivalent to the best treatment accorded to any ‘comparable’ domestic or other foreign market participant. Under this approach, non-discrimination has a strong liberalizing effect and far reaching consequences for the regulatory autonomy of the contracting parties, in particular if additionally the comparator clause is interpreted widely.
Here, I would suggest that the "best treatment" standard is not really a "non-discrimination" standard, because no discrimination against foreign actors exists. He describes a situation where "a measure negatively affects only 1 out of 100 foreign market actors" and "99 out of 100 domestic market actors are also negatively affected." If anything, there is discrimination against domestic actors (disparate impact at least). But I don't see how you can look at this situation and say there is discrimination against foreign actors (barring some evidence of specific intent). I'm not sure Diebold would agree with me on this point, although later he says some things that perhaps indicate that he would:
Overlap between non-discrimination and non-restriction
Depending on how each element of the non-discrimination obligation is construed, the result may be that non-discrimination overlaps with the more integrative principle of non-restriction (Beschränkungsverbot). The legal concept of non-restriction goes much further in trade liberalization than the principle of non-discrimination.
Some commentators suggest that de facto discrimination in WTO law should also be interpreted such that any measure which is more burdensome than necessary for foreign services and suppliers should qualify as discrimination in violation of Articles II or XVII GATS. Under this approach, the elements of ‘less favourable treatment’ and ‘likeness’ would have to be replaced by a test of necessity and proportionality. Yet, another way of assimilating non-discrimination to non-restriction would be to combine the ‘best of the best treatment’ approach for ‘less favourable treatment’ with a broad economic or interpretation of the comparator clause. As it would be almost always possible to determine at least one distant competitor receiving more favourable treatment, the principle of non-discrimination would in essence be transformed into an obligation of non-restriction.
I agree that there is some "overlap" here, but I think it's also the case that, to a great extent, "non-discrimination" and "non-restriction" just cover different issues. Whether it is "non-restriction," or "best treatment," or "necessity," these principles all go beyond non-discrimination. It's great to have a debate on what principles should be included in IEL agreements. But I think it confuses the issue to refer to these principles in the context of non-discrimination. Going back to his situation where 1 out of a 100 foreign actors are negatively affected, it may be that this particular actor has received unfair, unreasonable, unnecessary, overly restrictive, or just generally bad treatment. But that's something different from discrimination.
Of course, someone might point out that I'm assuming non-discrimination is "nationality" based. I have to concede that this is true. My only response is that if trade and investment treaties are to contain a general non-discrimination principle that is not tied to nationality, I think that should be discussed, debated and written down in the text clearly. As things stand now, it seems pretty clear to me that the text of all the treaties I have seen is about "nationality" based discrimination. On the other hand, I recognize that not everyone sees it the same way, and it may be pretty clear to them that non-discrimination is not just about nationality.
Now he gets into issues of intent/purpose/motivation:
Fragmented standards and relevance of ‘regulatory purpose’
The main objective of non-discrimination obligations in international economic law is to outlaw measures which are specifically designed to protect the domestic market from foreign competition. ...
I'm pretty sympathetic to this view. I'm not sure everyone else is, though. My sense is that many, if not most, people get nervous about the use of intent here.
As to how intent should be included, he notes:
On this point, I would argue that the word "treatment" can be used to address issues of intent. The following aren't necessarily the best definitions of that term, but they seem like reasonably good ones, and, more importantly, they are online, so I can link to them:From a practical and pragmatic perspective it seems indifferent whether the regulatory purpose is considered under the comparator clause, the ‘less favourable treatment’ element or as a distinct and separate element. However, from a doctrinal and systemic angle it would be welcomed if the jurisprudence, or preferably the treaties themselves, would clarify whether and under which title the purpose of an allegedly discriminatory measure may be analysed. Such transparency would enhance legal security and facilitate the parties to a dispute to build their legal arguments. For clarity and structural reasons, the regulatory purpose should ideally be considered as its own legal element. However, due to the lack of a textual basis, adjudicating bodies mostly rely on the comparator clause or to a lesser extent on the element of ‘less favourable treatment’.
The Free Dictionary: Treatment - The act, manner, or method of handling or dealing with someone or something
Cambridge Dictionaries Online: Treat - to behave toward (someone) or deal with (something) in a particular way
Granted, neither definition talks explicitly about the intent of the actor. Nonetheless, I see in both of them the idea that the motives of the actor are relevant. I want to avoid taking dictionary definitions too far, so I'm not going to look up all the words in these definitions (though I admit I was tempted!). But the references to the "manner" or "method" of "handling or dealing with," and "behave toward" or "deal with" in "a particular way," all suggest to me that the thinking of the actor is part of the concept of "treatment."
Others take a different view of this, of course. Many people who want to include intent think it goes best in "likeness." Along these lines, this is from para. 118 of the Corn Products NAFTA Chapter 11 decision: "there is a close relationship between whether the State intentionally discriminated on grounds of nationality and the test of like circumstances."
Finally, here is Diebold's proposal for non-discrimination:A FLEXIBLE FACTOR-BASED STANDARD OF NON-DISCRIMINATION
Considering that the interpretation of non-discrimination obligations in international
economic law lacks coherence, which in turn creates legal uncertainty for the contracting
parties, individuals and parties to a dispute, the present section attempts to develop a factor
based approach to non-discrimination. Pauwelyn argued that the non-discrimination analysis
should treat ‘likeness’ as a mere threshold question and focus more specifically on ‘less
favourable treatment’ as the substantive test, taking into account a mix of elements. Section
III takes up this theory, suggesting that the entire non-discrimination analysis could be
viewed as a threshold question. In other words, ‘less favourable treatment’, ‘likeness’ and
other elements such as ‘so as to afford protection’ or ‘regulatory purpose’ should not be
incorporated in non-discrimination provisions as strict legal conditions which must be proven
by the complainant or the respondent pursuant to the applicable standard of review. Instead,
all the relevant elements could be viewed as soft-factors to be weighed and balanced in order
to come to an overall conclusion on whether or not a measure is discriminatory and thus
illegal.
I read this as saying that all the various elements should be considered, in terms of their degree, and then weighed and balanced against each other. Thus, for example, the degree of disparate impact (e.g., 95% vs. 60% disproportionate disadvantage on imports), the amount of evidence showing discriminatory intent (e.g., statements of legislators, lack of connection between the measure and its purported objective), and the degree of competition between the products/services/investors at issue, should be taken into account.
On this point, I think the tendency to separate out the various elements can sometimes obscure the big picture. I see the "flexibility" aspect of his proposal as designed to address this problem.
Here's a little more detail on how he thinks the analysis should go:
The first analytical step should be to determine whether the measure differentiates directly on the basis of origin or on the basis of other criteria. All forms of de jure differentiations which affect the competitive opportunities to the detriment of foreign market participants constitute a strong factor pointing towards illegal discrimination. However, the respondent must have the opportunity to justify the measure by proving its legitimate policy objective and a strong nexus between the measure and the objective under scrutiny.
...
If a complainant challenges a measure which differentiates on a basis other than origin (de facto discrimination), then the analysis needs to focus strongly on the effect of the measure in a specific market situation.
...
Future non-discrimination provisions in international economic agreements should follow these examples, prompting the adjudicating bodies to assess the approximate extent of the competitive relationship by focusing on the economic theory of demand substitutability.
...Once the approximate extent of the competitive relationship between the foreign and domestic market entity has been established, the question becomes whether there is less favourable treatment of the foreign entities. For this purpose, the adjudicating bodies need to assess the effect of the measure, both qualitatively and quantitatively.
...
A final element which may be taken into consideration is the regulatory purpose of the measure under scrutiny. Importantly, the regulatory purpose should have its own value and should not be incorporated into the analysis of the previous elements.
This last part was a little hard to summarize, so perhaps at this point, I should just suggest checking out the paper itself.
