by Michael Dorf
In my latest Verdict column I warn of a possible danger in King v. Burwell, the latest SCOTUS challenge to Obamacare: I explain that even if a majority of the Court rejects the petitioners' argument that subsidies should not be permitted on a federally run health insurance exchange, the Affordable Care Act could still go up in flames depending on how CJ Roberts votes. Suppose that CJ Roberts concludes that subsidies are unavailable but that Justice Kennedy splits with the other conservatives to say that they are available. The law survives 5-4, right? Maybe not, I argue in the column. Perhaps Justice Kennedy will adhere to his view from NFIB v. Sebelius that the individual mandate is unconstitutional and non-severable from the rest of the ACA, including the subsidies. Confused? Go read the column. It explains the murky law governing when a Justice adheres to his prior dissenting vote.
The column addresses the question of how each of the four dissenting NFIB Justices should decide whether to adhere to his prior vote. But there is a further question about how to tally those votes: whether the Court counts votes by issue or by outcome. To understand the difference, let’s suppose that after the oral argument in King, the Justices go around the table and state their respective views as follows: Six Justices (Kennedy, Ginsburg, Breyer, Alito, Sotomayor, and Kagan) say that subsidies are permitted on federal exchanges; three Justices (Roberts, Scalia, and Thomas) say that subsidies are not permitted on federal exchanges; and the four Justices who previously voted to invalidate the entire ACA (Scalia, Kennedy, Thomas, and Alito) reiterate that view because they continue to think that it is inextricably intertwined with the individual mandate, which they believe is beyond the power of Congress.
What is the bottom line? If we look at the matter issue by issue, the ACA survives. The Court rules 6-3 to permit subsidies on federal exchanges; and only 4 Justices express interest in overruling the 2012 decision sustaining the individual mandate against the constitutional challenge, so the law remains constitutional.
But if we look at the bottom lines (in my admittedly hypothetical version of the case), we get a different result. Only 4 Justices think that the law both falls within the power of Congress and permits subsidies on federal exchanges. If the question is whether the subsidies survive, outcome voting gives a negative answer: two Justices (Kennedy and Alito) think that the law permits subsidies on federal exchanges but is unconstitutional; one Justice (Roberts) thinks the law is constitutional but forbids subsidies on federal exchanges; and two Justices (Scalia and Thomas) think that the law forbids subsidies on federal exchanges and is also unconstitutional.
So which is it? Issue voting or outcome voting? Somewhat surprisingly, the law gives no clear answer. The matter appears to be left to the discretion of individual Justices. Sometimes a Justice who would be inclined to overrule a prior decision if faced squarely with that possibility will nonetheless accept the prior case’s authority for purposes of deciding the new case. In other words, he or she will cast her bottom-line vote—to reverse or to affirm the lower court—based on her tally of the issue-by-issue voting.
But not always. When the Court fractures so that there is no majority opinion that commands five votes, outcome voting usually prevails, so that the bottom line is constructed from varying, even conflicting rationales. The 1949 case of National Mutual Ins. Co. v. Tidewater Transfer Co. is an arresting example. There a majority of Justices rejected the claim that Congress could vest jurisdiction in the federal courts over cases besides those listed in the Constitution’s Article III. A majority also rejected the proposition that the District of Columbia counted as a state for purposes of the provision of Article III permitting jurisdiction over state law cases involving citizens of different states. Yet because those conclusions were reached by different majorities, the bottom line was that Congress was permitted to authorize jurisdiction over state law cases between citizens of the District and citizens of a state.
As I have observed before, there is no all-purpose answer to the question of whether and when to use outcome voting versus issue voting. In that earlier post, I noted that, given how frequently judges disagree about all sorts of matters, it's a little surprising that the outcome-versus-issue voting question doesn't arise more frequently. I speculated that judges are pragmatists who must reach a judgment, and that this imperative drives them to submerge these matters. But at least four Justices thought that King was a sufficiently weighty case to grant cert despite the absence of a split (following the en banc DC Circuit decision to vacate its contrary opinion on the same issue). That's either a sign that at least four Justices are not thinking pragmatically or, worse, that they are--and that they have concluded that this time around they may be able to capture a fifth vote to kill the ACA.