In his post below, my co-blogger Ilya writes:
[F]or those who do believe that the legitimacy of judicial review depends at least in large part on the degree to which it has majoritarian support, the Court’s very high approval ratings relative to those of the president and Congress are not easy to dismiss. As I argued in my earlier post, they undercut arguments such as Orin’s, which hold that judicial review must be strictly limited because legislative enactments have a degree of popular “consent” that the courts lack.
Well, I wouldn’t want to “dismiss” any arguments, but I do disagree with Ilya about the relevance of the Supreme Court’s current approval ratings to the debate he and I have been having.
The key problem is that modern U.S. Supreme Court has acted as a highly majoritarian institution. No matter what power the Supreme Court has in theory, it has very rarely invalidated politically popular laws. This was particularly true in the last decade, when Justice O’Connor was the “swing vote” in most cases. As my colleague Jeffrey Rosen has pointed out, Justice O’Connor’s approach was very majoritarian: for the most part she kept the Court’s decisions exactly on track with American public opinion. When the Supreme Court effectively “follows the election returns,” it is bound to be pretty popular.
“But wait,” you’re thinking, “how can the Supreme Court strike down laws and yet also be popular if the people aren’t somehow ‘consenting’ to their laws being invalidated?” There are several reasons, but one is that public opinion polls are national whereas most laws struck down by the Supreme Court are (and traditionally have been) state or local. If a law is very popular in one state or region but unpopular nationwide, striking down the law will both thwart state and local preferences and be popular in national polls. To my mind, this doesn’t show that “the people” effectively “consented” to their law being struck down. Rather, it mostly shows that in a federal system, state and national opinion can and often do diverge.