By Mike Dorf
Judge Reinhardt's 2-1 majority opinion in Perry v. Brown holds that California's Prop 8 is unconstitutional on California-specific grounds. Relying heavily on the Supreme Court's 1996 decision in Romer v. Evans, the opinion emphasizes two facts about the case that, taken together, make it applicable only to California: A) California law extends to same-sex couples all of the legal incidents of marriage except the use of the word; and B) Prop 8 (prospectively) took away the right to marry that same-sex couples had previously enjoyed, rather than failed to extend a right that had never previously extended to same-sex couples. The opinion cites principles of judicial restraint for deciding the case on these relatively narrow grounds, and it also cannot have been lost on the majority that deciding the case in a way that only applies to California makes it somewhat less likely that the SCOTUS will see it as worthy of review.
In a follow-up post next week I may have something to say about the likelihood that the SCOTUS will take the case, but for now I want to focus on the substance of the 9th Circuit's decision to decide narrowly.
I'll begin by noting that each of the purportedly narrowing criteria creates the possibility of perverse incentives. A) The opinion repeatedly rejects asserted justifications for Prop 8 on the ground that the justifications, if credited, would mean that same-sex couples should not be given the operative rights that they have been given through its domestic partnership law, but that California in fact provides for such rights. Therefore, a state electorate might worry that enacting a domestic partnership law would lead to the courts requiring marriage, and instead give same-sex couples no or limited rights. B) Likewise, a state electorate or court that is considering experimenting with granting a right to marry might not do so for fear that it could never change its mind. Judge Reinhardt rejects these arguments as inconsistent with Romer and equal protection more broadly. I think he's right to reject the perverse incentive argument with respect to A) but that B) presents a trickier question.
As Judge Reinhardt's opinion correctly explains, it is often true that government violates the Constitution by doing more than it had to do but also less than it had to do. To use his example, there is no constitutional obligation for the government to provide food stamps for the poor, but if the government does provide food stamps for some of the poor, it cannot then deny food stamps to others of the poor on grounds that deny equal protection. This is simply how equal protection works. So there is nothing inherently perverse about move A).
Move B), however, is more problematic. Judge Reinhardt's opinion says that the withdrawal of a previously granted right denies equal protection where the government lacks a rational basis for withdrawing the right. That much strikes me as correct -- and we can imagine circumstances in which it would be rational not to extend a right but irrational to take it away. Suppose that the state had not previously extended a right to use some new drug to redheads because of a reasonable but unproven fear that the gene for redheadedness also made people especially vulnerable to some lethal side effect of the drug. Failure to extend the right to use the drug to redheads would not be irrational. However, if redheads had the right to use the drug for a year and none of them succumbed to the side effect, then it would be irrational to take the right away because of the now-disproved fear (assuming that the side effect was expected to manifest quickly). Still, in this example the important distinction is not between failing-to-extend and taking-away; the real difference is that we evaluate the decision differently based on different knowledge of the relevant facts.
Here's a better category of examples: Taking away a right can be unconstitutional where failing to extend it in the first place is not, because taking away may reflect animus to some group of people, whereas failing to extend may simply reflect indifference or legislative inertia. It's possible that this is all that Judge Reinhardt's opinion means to say about Prop 8: The taking-away under the particular circumstances in which Prop 8 was passed shows animus towards LGBT persons in a way that a mere failure to legalize same-sex marriage in the first place perhaps would not have. Perhaps the majority highlights the taking-away aspect of the case only as evidence for the conclusion that Prop 8 reflects animus and thus fails the Romer test. If so, that is defensible, in my view.
However, parts of the opinion appear to go further. In particular, in evaluating the justifications for the law, Judge Reinhardt relies on the not-extending/taking-away distinction. For example, the Prop 8 sponsors argued that reserving the word marriage for same-sex couples was justified because opposite-sex couples but not same-sex couples are capable of accidental reproduction, and thus opposite-sex couples but not same-sex couples need the enticement of the special status of marriage to induce them to wed. Judge Reinhardt says that this rationale might be a rational basis for not granting same-sex couples a right to marry in the first place but is not a rational basis for taking the right away.
That strikes me as highly contestable. Either the state is permitted to treat its bestowal of the word marriage as a limited resource or it is not. If it is, then why can't the state notice that it has bestowed the word marriage when it did not need to, and take it back?
The majority arguably reads Romer as standing for some sort of general proposition that it is unconstitutional for the government to bestow a legal right and then take it away, even though the government was under no obligation to bestow the legal right in the first place. But I don't think Romer says that. Romer is the third case in a trilogy of cases involving impermissible "animus" directed against identifiable groups. Notably, the middle case in the trilogy, Cleburne v. Cleburne Living Center, involves a failure to extend a legal right, not a taking away of an already-granted right. Yet it is of a piece with Romer.
So Romer is best read as not drawing any sharp or general-purpose distinction between taking away and not granting rights. There is, in other words, no freestanding constitutional endowment effect (except perhaps with respect to certain property rights). It's possible to read the 9th Circuit as recognizing a kind of endowment effect based on Romer, but the reasoning is sounder if one understands the references to taking away as simply evidence of animus.