Kansas law provides that unless a capital sentencing jury concludes that the mitigating factors that apply to the defendant's crime outweigh the aggravating factors, it must sentence the defendant to death. The Kansas Supreme Court held that this law violates the Eighth and Fourteenth Amendments because it "impermissibly mandates the death penalty when the jury finds that the mitigating and aggravating circumstances are in equipoise." On June 26, in Kansas v. Marsh, the Supreme Court reversed in a 5 to 4 opinion by Justice Thomas.
The genealogy of this case starts with Furman v. Georgia, in 1972, in which the Court invalidated all existing death penalty statutes, and Gregg v. Georgia, in 1976, in which it upheld three new post-Furman capital sentencing laws. The received wisdom is that Furman outlawed the "arbitrary" imposition and execution of death sentences, and Gregg began the apparently endless process of describing the requirements for the "non-arbitrary"-and hence constitutional-use of capital punishment. The issue in Marsh was whether the Kansas death sentencing scheme violated the large, complex, and messy body of constitutional case law applying and interpreting Furman and Gregg.
Justice Thomas, for the majority, addresses this in a straightforward manner. First, he finds that the Court has jurisdiction to review the Kansas Supreme Court's judgment because it was final on the issue of penalty and was not supported by an independent state ground. Then he concludes that reversal is required by the Court's 1990 decision in Walton v. Arizona. Finally, he also concludes that even if Walton does not bind the Court, reversal is required because the Kansas statute is consistent with Furman and Gregg. One may disagree with the outcome or the reasoning of the opinion, but there is nothing remarkable about it, either in tone or in content.
The concurring and dissenting opinions, however, are another matter entirely. Justice Stevens, in a solo dissent, argues that the Court should not have granted certiorari to review a state court decision that, at worst, extends a protection in the Bill of Rights beyond what the Court requires. This is an argument he has made and lost before. The real action is elsewhere: in a heated exchange between Justice Souter, dissenting for himself and Justices Stevens, Ginsburg and Breyer on one side, and Justice Scalia concurring alone on the other. I will discuss the content of this exchange first, and then return to the peculiar fact that it happened at all.
Most of Souter's dissent addresses the arguments that Walton or other prior decisions are not controlling, and that the Kansas statute is invalid under Furman and Gregg. But Souter ends on a different note:
Today, a new body of fact must be accounted for in deciding what, in practical terms, the Eighth Amendment guarantees should tolerate, for the period starting in 1989 has seen repeated exonerations of convicts under death sentences, in numbers never imagined before the development of DNA tests.
Souter goes on to discuss several studies of false convictions in capital cases, and concludes that given "the hazards of capital prosecution" a preference for death when "the evidence pro and con [is] in equipoise is obtuse by any moral or social measure."
Scalia responds in blistering sarcastic detail: Because the position taken by the dissent will inevitably provoke "sanctimonious criticism of America's death penalty" from Europeans, he feels compelled to "take the trouble to point out"-at length-that the dissent "has nothing substantial to support it" and "is willing to accept anybody's say-so" about crucial facts.
Scalia's separate opinions are not known for restraint, and this one is hardly an exception. Among others, various "professors" are conspicuous objects of his scorn, including yours truly. Souter's dissent relies in part on a 2005 article written by several students and me, on the subject of exonerations in the United States from 1989 through 2003. Scalia, in response, describes our criteria for classifying a case as an exoneration as "self-congratulatory," and faults us for "inflat[ing]" the number of false convictions. As evidence he points to two cases out of 340 in which he says the exonerated defendants were in fact guilty.
A charitable description of this criticism would be "careless." My co-authors and I specifically pointed out that a few of these hundreds of exonerated defendants may in fact have been guilty. That follows from our decision to define "exoneration" by reference to the legal outcome of the case: it requires an authoritative decision that a defendant who had been convicted of a crime is released from all legal consequences of that conviction because of new post-conviction evidence of innocence. We thought it best to use that definition rather than rely on our own remote third-hand judgments of guilt or innocence. (Scalia, on the other hand-perhaps because of his position as a federal appellate judge-seems comfortable determining the true facts of state criminal cases.) Apparently we are accused of publishing a study in which we did exactly what we said we were doing.
Scalia's two main points, however, are more general and more important:
• First, Scalia recycles a claim made by pro-death penalty advocates. According to Scalia, even if we assume many more exonerations than have been publicized, considering the huge total of all criminal convictions in the United States, "the error rate [is] .027 percent-or, to put it another way, a success rate of 99.973 percent." In other words, false convictions are a drop in the bucket, too ridiculously rare to worry about.
In fact, the true number of wrongful convictions is unknown and frustratingly unknowable. But the rate that Justice Scalia advocates is flat wrong and badly misleading.
To come up with a rate for an event you should divide the number of times it has happened by the number of times it might have happened. Probably the most familiar example for Americans is the batting average: the number of hits by a ballplayer, divided by his number of at-bats. If you ignore some hits, or exaggerate the number of at-bats, a player's average will look lower than it really is. Both of these problems occur here.
Consider this example: After the first Gulf War, Desert Storm, the Veterans Administration ("VA") reported an unusually high number of cases of Lou Gehrig's disease-amyotrophic lateral sclerosis-among Gulf War veterans. Let us assume that 200 such cases were reported (the real number, thankfully, is lower-so far). One could say that this is an insignificantly small number, 200 divided by the 26 million veterans living in the United States, or 0.0008%. Of course, that is absurd. Shouldn't it be 200 divided by 700,000-the number of Desert Storm veterans-a rate 37 times higher? Or perhaps 200 divided by the 100,000-the number of veterans believed to have been exposed to nerve agents-about 260 times higher?
Scalia makes this elementary mistake. He divides his hypothetical number of exonerations by all felony convictions in the United States. But 96% of known exonerations occurred in rape and murder cases, which together constitute less than one in fifty felony convictions. This is not surprising. DNA, the most powerful evidence of innocence, is almost entirely limited to rape cases. And the average time from conviction to exoneration is ten years, but in the great majority of felony convictions-for burglary, assault or drug offenses-the defendants, guilty or innocent, are freed long before that. So the Scalia ratio, if it has any meaning, has to be limited to rape and murder, and multiplied by a factor of about 50-for starts.
Then there is the question of the numerator: How many false convictions are there really? We don't know. To return to Desert Storm, the VA's current figure on Lou Gehrig's disease among veterans is hardly the last word. A government report in 2004 warned that the number of cases might increase sharply in years to come as those veterans age; many cases will never be detected because of poor diagnosis or death from another cause; and some that are correctly diagnosed may not be reported to the VA.
The same is true here. We do not know about many exonerations that have already occurred. In addition, there will be more exonerations among defendants now in prison, especially among those who are there for decades or for life, and many other false convictions will never come to light. For example, almost all rape exonerations depend on DNA evidence, but testable biological samples have been located in only about a quarter of the rape convictions in which they have been sought. In robbery cases, where mistakes are probably more numerous than for rapes, DNA is almost never available and there have been almost no exonerations.
While we have no meaningful overall estimate of the rate of false convictions, we do have some hints at the magnitude of the problem. Here are two examples:
In about 25% of pre-trial DNA tests performed by the FBI, crime scene samples do not match the suspect. Most of those suspects are cleared, but how many would have been falsely convicted before DNA testing existed? And how many similar suspects are still falsely convicted for crimes like robbery, where DNA evidence hardly ever exists? Let us hope it is rare, but the prospect is sobering.
Among the 7,529 defendants sentenced to death in the United States from 1973 through 2004, 117 or 1.6% have been exonerated. Among the half who have been on death row longest-those (like older Lou Gehrig's disease patients) who have been incarcerated long enough to have had a real chance to be exonerated-2.4% have been freed. The true rate of false capital convictions, including convictions of capital defendants who are not sentenced to death, is probably considerably higher.
Is that a lot or a little? That depends on your point of view. If as few as 1% of jets crashed on takeoff, we would shut down every airline in the country. The good news is that the great majority of convicted defendants in America are guilty; the bad news is that a substantial number are not.
• Second, Scalia argues that the only issue that might matter to the status of the death penalty is the execution of innocent defendants. Other false convictions, capital and non-capital, are immaterial. On that issue (which the dissent never discusses) Scalia's view is clear: In recent American history, there has not been "a single case-not one-in which it is clear that a person was executed for a crime he did not commit. If such an event had occurred in recent years, we would not have to hunt for it; the innocent's name would be shouted from the rooftops."
I don't know about rooftops, but what about the Chicago Tribune and ABC News? Starting on June 23 of this year these two major news media ran overlapping series of stories based on a joint investigation of the case of Carlos DeLuna. DeLuna was executed in 1989 for stabbing a convenience store clerk to death in 1983 in Corpus Christi, Texas. He claimed that he had been misidentified at a hurried on-the-scene identification, and that the real killer was a man named Carlos Hernandez.
The prosecution argued to the jury that Hernandez was a "phantom." The Chicago Tribune, however, found that Carlos Hernandez (who since died in prison) was real and well known to law enforcement in Corpus Christie. In fact, one of DeLuna's prosecutors knew Hernandez well from an earlier homicide investigation. Hernandez looked like DeLuna, and-unlike DeLuna-had a long history of knife attacks similar to the 1983 convenience store killing. He also repeatedly told friends and relatives that he had committed the murder for which DeLuna was executed.
It is certainly understandable that Scalia missed this late-breaking account of an erroneous execution. The first report was just days before the Marsh decision, and the final installment came out after that opinion was filed. But the DeLuna story is only the most recent of several cases of false executions.
Cameron Willingham was sentenced to death for murder by arson in Corsicana, Texas in 1992. In December 2004, the Chicago Tribune reported that new scientific knowledge proves that the testimony by arson experts at Willingham's trial was worthless, and that there is in fact no evidence that the fire in question was caused by arson. Last March a panel of the nation's leading arson experts confirmed that conclusion. In a strikingly similar case, another Texan-Ernest Willis-was convicted on the same sort of evidence and sentenced to death for murder by arson in Pecos County, Texas in 1987. Willis was exonerated and freed in October 2004-eight months after Willingham was put to death, in February, 2004.
Larry Griffin was executed in Missouri in June 1995 for the drive-by shooting of a drug dealer in 1980. The only evidence connecting him to the crime was a witness who claimed to have seen Griffin firing shots from the murder car. This witness was a white career criminal with several felony charges pending against him. In July 2005, an investigation by the NAACP Legal Defense and Educational Fund revealed that the first police officer on the scene and the victim's sister both agreed that this supposed witness-who would have stood out in the all-black neighborhood-wasn't there when the shooting occurred. In addition, investigators located a second victim who was also injured in the shooting. He knew Griffin and says that Griffin was not in the car from which the shots were fired, but he was not called to testify at Griffin's trial. This evidence led the chief prosecutor of the City of St. Louis to reopen the investigation of the Griffin case, twenty-five years after the murder for which he was executed.
Ruben Cantu was executed in August 1993 for a robbery-murder that was committed in San Antonio in 1985, when he was 17. In November 2005, the Houston Chronicle reported that Cantu's co-defendant, who pled guilty to participating in the crime but did not testify at Cantu's trial, has signed an affidavit swearing that Cantu was not with him that night and had no role in the murder. More important, the only eyewitness who did testify-a second victim, who was shot nine times but survived-now says that the police pressured him to identify Cantu as the shooter, and that he did so even though he knew that Cantu was innocent.
None of these four has been officially exonerated. They may never be; there is no generally available procedure for securing an exoneration after execution. But the evidence that has come to light since they were put to death points overwhelmingly to their innocence. It might be comforting to believe that no innocent defendants have been executed recently in America, but the evidence says otherwise. We know it happens; we are leaning why and how; and now we know the names and faces of some of those who have been put to death for crimes they did not commit.
The most puzzling aspect of this conflict is not the content of Souter's dissent or the tone of Scalia's concurrence. It is the context: Why did this fight occur in the Marsh case? Scalia is right when he says that innocence was not an issue in that case. And he's also right when he says that the dissent-a statement by four members of the Court that recent death row exonerations raise general questions about the constitutionality of our use of the death penalty-will get a great deal of attention. Given his views on the constitutionality of capital punishment, Scalia's fervor is understandable, if not his treatment of facts.
So why did this fight erupt in Kansas v. Marsh? Perhaps the answer has to do with House v. Bell, an earlier capital case last term that really did concern innocence. The issues in House are dense and technical. Suffice it to say that the defendant, who had been sentenced to death for a rape-murder, assembled enough new evidence of innocence to persuade six judges on the Sixth Circuit sitting en banc that he should be released forthwith, and a seventh judge that he was entitled to a new trial-while a majority of eight judges saw no constitutional problem with putting him to death.
House arrived in Washington with an air of controversy. It was considered by the Court at five separate conferences before certiorari was granted. Ultimately, House won 5 to 3, with Justice Kennedy writing for the Court and Chief Justice Roberts dissenting. Unlike Marsh, however, the division on the Court was expressed as a polite legalistic disagreement over the degree of deference the Supreme Court should attach to a district court's decision under the indecipherable standard that the Court has laid out for federal courts to apply to such claims. There were no fireworks.
Was the dogfight in Marsh a leftover from House v. Bell, which (unlike Marsh) actually was about the danger of executing an innocent man? Or were these the opening shots in a new battle-a battle over whether the hundred-and-twenty-plus death row exonerations since 1973 amount to, as the dissent says, "a new body of fact [that] must be accounted for" in deciding what sort of death penalty the Constitution permits? Time will tell.