This week, Adam Liptak of the New York Times published this Sidebar piece  on death penalty stays at the Supreme Court entitled "Going to Court, but Not in Time to Live."

Liptak does great work.  I wanted to comment on this piece, only because it could leave the misimpression that it is possible for death penalty petitioners to secure enough votes to grant certiorari but not enough for a stay, so that the execution would take place before the Court heard the case.  If that were true, it would be deeply troubling; thankfully, it is not.

The piece notes the prospect of that scenario under the Court's unpublished practices, which require four votes to grant certiorari but a fifth vote for a stay of execution.  It cites as an example "the case of Luther J. Williams" who was executed notwithstanding that he received four votes for a stay.  The piece continues:  "Supreme Court math used to yield different results," citing a previous practice of a Justice providing a courtesy fifth vote ""when four members have voted' to hear an appeal." 

Liptak also reports the following exchange at John Roberts' confirmation hearings:

Judge John G. Roberts Jr. was asked what he would do "if you had four other justices now voting for a stay of execution?"

"Do you feel, as chief, you should do the courtesy," Senator Patrick J. Leahy, Democrat of Vermont, asked, "and kick in the fifth one?"

"I don't want to commit to pursue a particular practice," Judge Roberts said.  "But it obviously makes great sense."

"You don't want to moot the case by not staying the sentence," he added.


Given recent experience at the Court, Liptak says that, though it is not clear, "it does not seem that Chief Justice Roberts has consistently adopted the practice he had tentatively endorsed."

No doubt in part because the Court’s own practices in this area are opaque, I’m afraid that the Liptak piece doesn't capture the nuance of the distinction it hopes to draw between votes to grant certiorari and votes to grant a stay.  The practice in the later Rehnquist years "“ granting a courtesy fifth vote for a stay of execution when there are four votes to grant certiorari "“ remains fully in effect under Chief Justice Roberts, and that practice is in fact completely consistent with his testimony.

The confusion arises from the fact that there regularly are four votes to grant a stay of execution at the Supreme Court, but not four votes to grant cert.  That seems somewhat counterintuitive.  Why would four Justices say that an execution should be halted if they don't want to hear the case?  Liptak gives one reason:  Justices Stevens and Ginsburg will always vote to grant a stay for procedural reasons when the normal period for petitioning for certiorari on federal habeas corpus has not expired.  Here are three different examples.  Justices Souter and Breyer do not apply as strong a rule, but they too seem to vote to stay execution when presented with potentially meritorious cert petitions.  Here are three different examples, and here is one in which only Breyer joined Stevens and Ginsburg.

But "“ and here is the rub "“ when Justices Stevens, Souter, Ginsburg, and Breyer vote to grant a stay it does not necessarily mean that they would grant certiorari.  They just have concluded that the execution should not be carried out while the petition is pending and the Court is considering it.  Those Justices can "“ and sometimes do "“ go a step further and vote to hear the case.  (Even if only the stay application (rather than a petition for cert.) is pending, they can deem the application to be a petition and vote to grant it.)  If four do decide to hear the case, then under current practice a courtesy fifth vote will be provided for a stay of execution.  That is the reason that for several years you have not seen any cases in which the Court publishes an Order stating that an application for a stay of execution is denied but four Justices would grant cert.

All of this returns us to the exchange as reported by Liptak from the Roberts hearings.  It is true that Senator Leahy asked if Roberts would provide a fifth vote for a stay "if you had four other justices now voting for a stay of execution."  But Roberts answered Leahy's question more carefully (and absolutely correctly) than Liptak recognizes (the underlining below is the part of the response quoted in the article; the italics are the important part of the answer that the article does not quote in which the Chief Justice draws the relevant distinction):

It's an issue that I'm familiar with.  I do know it arose.  And I thought the common practice "“ the current practice "“ was that, if there are four votes to grant cert, that the court would grant the stay even though that does require the fifth vote, so that you don't have a situation.

. . . .

I think that practice makes a lot of sense.  I don't want to commit to pursue a particular practice in an area that I'll obviously have to look at in the future.  But it obviously makes great sense that, if you have four to grant and that's the rule, that you will consider an issue if there are four to grant, you don't want to moot the case by not staying the sentence.


So, in the exchange, Roberts corrects Senator Leahy's understanding of the Court's practice (i.e., he explains that the courtesy fifth actually is triggered by four votes for cert rather than for a stay) and says that he thinks that's sensible.  And the Court has in fact continued to follow that practice.

Posted in Everything Else