We're Not Executing the Innocent

By Paul G. Cassell
Copyright 2000 Wall Street Journal
June 17, 2000

On Monday avowed opponents of the death penalty caught the attention of Al Gore among others when they released a report purporting to demonstrate that the nation's capital punishment system is "collapsing under the weight of its own mistakes." Contrary to the headlines written by some gullible editors, however, the report proves nothing of the sort.

At one level, the report is a dog-bites-man story. It is well known that the Supreme Court has mandated a system of super due process for the death penalty. An obvious consequence of this extraordinary caution is that capital sentences are more likely to be reversed than lesser sentences are. The widely trumpeted statistic in the report -- the 68% "error rate" in capital cases -- might accordingly be viewed as a reassuring sign of the judiciary's circumspection before imposing the ultimate sanction.

Deceptive Factoids

The 68% factoid, however, is quite deceptive. For starters, it has nothing to do with "wrong man" mistakes -- that is, cases in which an innocent person is convicted for a murder he did not commit. Indeed, missing from the media coverage was the most critical statistic: After reviewing 23 years of capital sentences, the study's authors (like other researchers) were unable to find a single case in which an innocent person was executed. Thus, the most important error rate -- the rate of mistaken executions -- is zero.

What, then, does the 68% "error rate" mean? It turns out to include any reversal of a capital sentence at any stage by appellate courts -- even if those courts ultimately uphold the capital sentence. If an appellate court asks for additional findings from the trial court, the trial court complies, and the appellate court then affirms the capital sentence, the report finds not extraordinary due process but a mistake. Under such curious scorekeeping, the report can list 64 Florida postconviction cases as involving "serious errors," even though more than one-third of these cases ultimately resulted in a reimposed death sentence, and in not one of the Florida cases did a court ultimately overturn the murder conviction.

To add to this legerdemain, the study skews its sample with cases that are several decades old. The report skips the most recent five years of cases, with the study period ostensibly covering 1973 to 1995. Even within that period, the report includes only cases that have been completely reviewed by state appellate courts. Eschewing pending cases knocks out one-fifth of the cases originally decided within that period, leaving a residual skewed toward the 1980s and even the 1970s.

During that period, the Supreme Court handed down a welter of decisions setting constitutional procedures for capital cases. In 1972 the court struck down all capital sentences in the country as involving too much discretion. When California, New York, North Carolina and other states responded with mandatory capital-punishment statutes, the court in 1976 struck these down as too rigid. The several hundred capital sentences invalidated as a result of these two cases inflate the report's error totals. These decades-old reversals have no relevance to contemporary death-penalty issues. Studies focusing on more recent trends, such as a 1995 analysis by the Criminal Justice Legal Foundation, found that reversal rates have declined sharply as the law has settled.

The simplistic assumption underlying the report is that courts with the most reversals are the doing the best job of "error detection." Yet courts can find errors where none exist. About half of the report's data on California's 87% "error rate" comes from the tenure of former Chief Justice Rose Bird, whose keen eye found grounds for reversing nearly every one of the dozens of capital appeals brought to her court in the 1970s and early 1980s. Voters in 1986 threw out Bird and two of her like-minded colleagues, who had reversed at least 18 California death sentences for a purportedly defective jury instruction that the California Supreme Court has since authoritatively approved.

The report also relies on newspaper articles and secondhand sources for factual assertions to an extent not ordinarily found in academic research. This approach produces some jarring mistakes. To cite one example, the study claims William Thompson's death sentence was set aside and a lesser sentence imposed. Not true. Thompson remains on death row in Florida today for beating Sally Ivester with a chain belt, ramming a chair leg and nightstick into her vagina and torturing her with lit cigarettes (among other depravities) before leaving her to bleed to death.

These obvious flaws in the report have gone largely unreported. The report was distributed to selected print and broadcast media nearly a week in advance of Monday's embargo date. This gave ample time to orchestrate favorable media publicity, which conveniently broke 24 hours before the Senate Judiciary Committee began hearings on capital-sentencing issues.

The report continues what has thus far been a glaringly one-sided national discussion of the risk of error in capital cases. Astonishingly, this debate has arisen when, contrary to urban legend, there is no credible example of any innocent person executed in this country under the modern death-penalty system. On the other hand, innocent people undoubtedly have died because of our mistakes in failing to execute.

Real Mistakes

Collen Reed, among many others, deserves to be remembered in any discussion of our error rates. She was kidnapped, raped, tortured and finally murdered by Kenneth McDuff during the Christmas holidays in 1991. She would be alive today if McDuff had not narrowly escaped execution three times for two 1966 murders. His life was spared when the Supreme Court set aside death penalties in 1972, and he was paroled in 1989 because of prison overcrowding in Texas. After McDuff's release, Reed and at least eight other women died at his hands. Gov. George W. Bush approved McDuff's execution in 1998.

While no study has precisely quantified the risk from mistakenly failing to execute justly convicted murderers, it is undisputed that we extend extraordinarily generosity to murderers. According to the National Center for Policy Analysis, the average sentence for murder and non-negligent manslaughter is less than six years. The Bureau of Justice Statistics has found that of 52,000 inmates serving time for homicide, more than 800 had previously been convicted of murder. That sounds like a system collapsing under the weight of its own mistakes -- and innocent people dying as a result.

Mr. Cassell is a professor of law at the University of Utah.