I'm going to try to come back to this for more detailed analysis when I have more time, but at first glance I really like this non-discrimination paper (from the SIEL conference going on right now) by Nicolas Diebold. Here's a brief excerpt:
... Pauwelyn argued that the non-discrimination analysis should treat ‘likeness’ as a mere threshold question and focus on ‘less favourable treatment’ as a legal element. This article takes this theory one step further, 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 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. ...
It may be counterintuitive, but I think it's possible that non-discrimination analysis could be made more precise by adding more flexibility. The individual elements could be assessed, then weighed and balanced against each other. Here's a question we asked in the non-discrimination chapter of our textbook, along these same lines:
A number of different elements have been proposed as the basis for identifying discrimination. Putting aside the issue of the appropriate part of the provision under which they should be considered, what do you think of a standard that would weigh and balance all of the relevant considerations, as follows. A non-discrimination standard would take into account: (1) the existence of, and degree of, disparate impact on imports; (2) the various subjective intents of the government officials involved in the enactment of the measure; (3) the objective intent of the measure (including an examination of whether the means-ends relationship of the measure and its stated goals).
************************************
Many other papers from the SIEL conference are available here.
