by Michael Dorf
Imagine a conservative judge on a panel on one of the courts of appeals in a circuit that has not yet ruled on the constitutionality of state laws forbidding same-sex marriage (SSM). Let's call him Judge Hetero. Judge Hetero has considered all of the arguments, and his best legal judgment--informed as a judge's legal judgment inevitably is, by his values--is that there is no constitutional right to SSM. Hetero admits that the question is close in light of United States v. Windsor, but he thinks that CJ Roberts was ultimately correct in reading that case as a federalism decision. I know most of my readers do not think this; I don't think this; I think the opposite; but at least a few judges think this, and I'm asking you to assume for reasons that shall become clear that Hetero is one of these judges.
If Hetero's court had been required to rule on the validity of a state SSM ban last week, he would have voted to uphold the ban. Now the question: Assuming Hetero is acting in perfectly good faith, respecting precedent, and not playing any strategic games, but that he remains as conservative on this issue as he was last week, should he change his mind in light of the cert denials? Somewhat surprisingly, the answer is not clear.
Now, in one obvious sense the cert denials make no difference. As the Supreme Court has repeatedly noted, a cert denial sets no precedent. So formally, the cert denials are irrelevant and a judge who was inclined to rule against a right to SSM last week should remain so inclined this week.
But the cert denials may be relevant in a different sense: They enable the appeals court judges to predict with near certainty that the Supreme Court (at least absent any retirements or deaths) would vote to uphold a right to SSM. As a judge on what the Constitution calls an "inferior Court," Hetero might well think that his job is to decide cases in a way that will not result in being reversed by his superiors on the Supreme Court. Indeed, it can be and has been argued that the very obligation of vertical stare decisis is an outgrowth of the ability of a superior court to reverse an inferior one. So why waste everyone's time and money by issuing a decision that Hetero has extremely good reason to believe will be reversed?
The answer depends on how we conceptualize the role of lower court judge. Is it the job of a lower court judge to predict how a higher court will decide an issue? In one context, the Supreme Court has said no. In 1989, in Rodriguez de Quijas v. Shearson/American Express, Justice Kennedy said this for the majority: "If a precedent of this Court has direct application in a case, yet appears to rest on reasons rejected in some other line of decisions, the Court of Appeals should follow the case which directly controls, leaving to this Court the prerogative of overruling its own decisions." Suppose Hetero thinks that the Supreme Court's 1972 order summarily dismissing the appeal in Baker v. Nelson as presenting no substantial federal question was a judgment on the merits that there is no constitutional right to SSM--as briefs for opponents of SSM argue. Then Hetero concludes, based on Rodriguez de Quijas, that he should ignore the signs that the Supreme Court would now rule otherwise, because they merely undermine but do not expressly overrule Baker.
However, nearly all lower courts to have considered the issue find that the Rodriguez de Quijas rule either does not apply to a per curiam appeal dismissal like Baker or that the threshold for superseding such a bare ruling is lower for that kind of case. Suppose Hetero does not think that Baker decides the issue but his best judgment is nonetheless that there is no right to SSM. Now should Hetero sublimate his best judgment to his prediction of how the Supreme Court will decide the case?
I wrote an article in the 1995 UCLA Law Review (downloadable here) in which I argued that lower court judges generally should not conceive of their role as predicting the rulings of higher courts. Predicting how other judges will rule on an issue, I recognized, is accepted in special circumstances, such as when a single Justice must decide whether to grant a stay and acts as a proxy for the full Court or, more controversially in my view, when a federal district court sitting in diversity must resolve an unresolved question of state law. However, I contended that as a general matter, rule-of-law values concerning the impersonality of legal principles obligate lower court judges to apply their own best judgment about how to apply the law--including, of course, all applicable precedents of higher courts.
I continue to think that lower court judges generally should not reject their best judgment about legal principles in favor of a prediction of how a higher court will rule. Accordingly, my answer would be that Judge Hetero should stick to his guns and rule against SSM. If the Supreme Court wants to reverse him based on a different (equally sincere) view of the law, that's its business, not his.
However, I recognize that the issue of whether and when courts should predict rather than use their own best judgment is not settled. Almost exactly simultaneously with my UCLA article, Evan Caminker published an article in the Texas Law Review (no general access version but available via Hein, WestLaw, or Lexis to those with subscriptions) in which he argued for more or less the opposite conclusion. (Caminker and I did not learn of our respective undertakings until we were both nearly finished with our papers, just in time to cross-cite one another but not in time for either of us to address the other's arguments as such in an extended fashion.)
Finally, note that the question whether a lower court judge should predict or simply decide cases according to her own best judgment does not have a clear ideological valence. When the higher court is to the right of the lower court judge, predicting will skew to the right, while best judgment will skew to the left, and vice-versa when the higher court is to the left of the lower court judge. My hypothetical example with SSM falls into the latter category but there is no systematic bias to the choice.