Having totally predicted the outcome of today’s ObamaCare decision, I feel obliged to try to guess the vote count on tomorrow’s (or next week’s?) gay-marriage outcome too. In some ways this is an easier call than King v. Burwell, in some ways it’s harder. It’s easier in the sense that the outcome isn’t in doubt. There was, at least, a chance (according to some people, not me) that Roberts and Kennedy would side with the other conservatives and strike down federal O-Care subsidies. To predict King correctly, you had to guess the outcome before you guessed the votes. The gay-marriage case is easier insofar as the outcome isn’t in doubt. Eight votes are all but assured: Kennedy and the four liberals will form a majority in favor of legalizing gay marriage and Scalia, Thomas, and Alito will dissent.

Roberts is less predictable on gay marriage, though, than he was (in my opinion) on O-Care. The chances that he’d undercut the health-care law now after shocking the country by upholding it in 2012 seemed close to zero to me. As chief, he cares about the Court’s reputation as a non-political institution; if he’d switched and voted against ObamaCare today, lefties would have accused him of bowing to political pressure from righties after his infamous ruling on the mandate. Meanwhile, actually striking down the subsidies would have accomplished nothing in practice since the feckless congressional GOP inevitably would have surrendered to O and restored them. Faced with those options, of doubling down or of being perceived as a weasel to no practical policy effect, Roberts chose door number one. Today’s betrayal hurts on a visceral level because this Republican appointee whom we all swooned over in 2005 stabbed us in the back by siding with Obama again, but in terms of its consequences for policy the decision was almost irrelevant. Subsidies were getting restored no matter what. It was just a matter of which Republican-controlled branch would be handed the knife to do the back-stabbing.

So Roberts, having already heavily invested his reputation in ObamaCare in 2012, kicked in a little more today. Predictable. He’s not similarly invested in the Court’s gay-rights jurisprudence, which makes the forthcoming gay-marriage ruling less predictable. In fact, Roberts dissented — on mainly procedural grounds — in the landmark Windsor case two years ago that struck down part of DOMA and required the federal government to follow state laws on marriage. Scalia predicted at the time that the core holding of Windsor, which criticized DOMA for attempting “to impose inequality [on gay couples], not for other reasons like governmental efficiency,” would apply just as well to state marriage laws, making it all but inevitable that the Court would eventually strike down those laws too as violations in the Fourteenth Amendment. That logic is highly likely to influence tomorrow’s opinion — but John Roberts, given a chance to endorse it two years ago, declined and dissented. Then again, as SSM supporters have noted, he also declined to sign onto Scalia’s dissent, which attacked the logic of Kennedy’s majority opinion. Which way does he go this time?

Most righties, disgusted by his health-care ruling this morning, will say, “With the left, of course. That’s what traitors do.” I’m not so sure. The strongest argument for why he might vote with the liberals is to promote the idea of certainty. If, as is claimed, he cares about perceptions of the Court’s legitimacy, a 6-3 pro-SSM decision joined by two Republican appointees is harder for skeptics to dismiss as Kennedy’s folly. It’ll also be perceived by future justices as a bit more solid precedentially than a 5-4 ruling might be. On the other hand, realistically, even a 5-4 decision in favor of gay marriage stands very little chance of being overturned later. Republicans won’t want to take up this issue again down the road given the drift in public opinion towards gay rights, no matter what GOP presidential candidates might tell you. And even a strong conservative appointee to the bench would think twice about reversing a pro-SSM ruling for fear of the Court being perceived as fickle on a hot button that touches on people’s intimate personal lives. (How many thousands of legal marriages would be jeopardized if the Court reversed itself later?) Even if Roberts dissented in tomorrow’s decision, he’s young enough that he’d probably still be on the Court if the ruling was revisited 10 or 20 years from now. He could always change his vote to “yes” then. Long story short, unless he’s been persuaded on the constitutional merits of SSM, there’s little reason for him to vote with Kennedy and the liberals this time apart from the institutional argument that a 6-3 ruling always enjoys more respect.

Or is there? The NYT considered this question two months ago and offered another reason why Roberts might flip to “yes”:

[A] Supreme Court ruling authorizing the states to restrict marriage to heterosexual couples would lead to enormous confusion. The court’s decision not to take appeals of earlier cases, or put other appeals of lower court decisions on hold, has induced many couples to marry. Surely, same-sex couples who married when marriage was legal under prior court rulings would remain so. But if the plaintiffs do not prevail, all of the states that recognize the right of same-sex couples to marry in light of lower federal court rulings could revert to their earlier state of affairs…

The consequences of such a ruling would be dramatic — and quite unsettling. Administrative agencies would have to shift policies (again), create new forms (again), and educate their bureaucracies on a variety of matters — all of which could be undone by a new legislative act or state constitutional amendment. Given the sheer number of economic rights, presumptions and obligations that travel under the “marriage” heading, denying gay couples access to a uniform institution, with all the complications, headaches and (possibly) litigation it could bring, seems like the antithesis of restraint.

Time for predictions, then. I think it’s thisclose to a coin flip on how Roberts goes, but I’m going to guess that he sides with Kennedy and the liberals again to form a 6-3 majority. The institutional arguments will weigh on him. So might the rhetoric coming from GOPers like Mike Huckabee, who’s now vowing openly that he’ll resist a SCOTUS ruling that legalizes gay marriage as president. (What exactly that would mean, given that the decision would apply to state marriage laws, not the federal one, is unclear.) If Roberts is worried about rhetoric promoting a constitutional crisis gaining traction on the right, a 6-3 decision will help counter it. Plus, who knows? Maybe he and Kennedy made a bargain on health care and gay marriage to be each other’s political cover. Roberts was going to be the lightning rod for righty criticism by siding with the left on ObamaCare; Kennedy’s going to be the lightning rod by siding with them on SSM. By standing with each other on both decisions, they each get to hide a bit in a (small) crowd. I think it’ll be 6-3 — but I’m much less confident of that than I was about the King case.