Document Type

Book Chapter

Publication Date



It is obvious that no anti-discrimination regime can stop at forbidding explicit discrimination of the relevant sort. If only explicit discrimination is forbidden, lawmakers who want to discriminate can hide their discriminatory intentions behind facially neutral classifications that are nonetheless chosen because they differentially burden the protected class. So, we must be prepared to invalidate some facially neutral laws that have "discriminatory effect" or, as American lawyers often call it, "disparate impact." On the other hand, we cannot possibly invalidate all laws which have a disparate impact on a protected class; many perfectly reasonable laws adopted for completely innocent purposes will have such disparate impact. So, some laws with disparate impact must be upheld, and some must be invalidated. The question is how to draw the line. (I have intentionally not used the phrase "de facto discrimination." This common phrase could be used simply as a synonym for "disparate impact" and could be similarly neutral on the ultimate issue of illegality. But my impression is that for many trade lawyers, "de facto discrimination" tends to mean laws with disparate impact that are in fact illegal, for whatever further reason is determinative. In contrast, in American constitutional law "de facto discrimination" normally connotes the absence of illegality. So the phrase seems best avoided.) As I say, the central question for this conference is how to decide which laws with disparate impact are illegal and which are not. But the first question that would occur to an American lawyer about the Arimani duckmeat law (see Appendix, p. 359) is a different one, namely, whether the law is preempted by federal statute or administrative regulation. I suspect a strong case could be made that it is so preempted, although I have not pursued the matter, because that is obviously not the issue primarily relevant to this comparative exercise. I shall discuss the status of the Arimani law under the dormant commerce clause, on the assumption that it is not preempted by statute or regulation. Still, the issue of statutory preemption is worth mentioning, since the possibilities for statutory preemption may affect our approach to issues about preemption by fundamental law. The United States has a central legislature with regulatory powers and a history of using them that go far beyond any analogue in the European Union or the WTO. It is possible that differences of this sort may either explain (empirically) or justify (normatively) different approaches to fundamental law preemption in the three systems.