Percent Error Rate Found in Death Case Study
Author calls serious problems 'epidemic'
The National Law Journal
June 9, 2000
past two decades, federal and state courts have overturned 68 percent of
the death sentences they have reviewed because of serious errors in their
trials, according to a new study. And in cases sent back for retrials,
82 percent of convicted capital defendants received new sentences that
were other than death -- including 7 percent who were found innocent.
"This study is designed to look at the system and how common mistakes are," says its author, law professor James S. Liebman, of Columbia University. "What we've found is it's not one state, it's not one case, but it's the system in which this kind of serious error is epidemic."
The study defines "serious error" as error that "substantially undermines the reliability of the guilt finding or death sentence imposed at trial" and that led a court to overturn the conviction, the sentence or both.
When the 68 percent overall error rate is broken down into its state and federal components, the study, which covers the period from 1973 to 1995, reports:
• Of 599 death sentences finally reviewed in their first federal habeas corpus petitions, federal courts overturned 40 percent, or 237, because of serious error.
• Of the 4,578 death sentences finally reviewed by state high courts on direct appeal, 41 percent, or 1,885, were thrown out because of serious error. When that rate is combined with sentences overturned at the next stage -- state post-conviction proceedings -- state courts found serious error in 47 percent of death sentences.
• Of 26 states that had at least one case move through the three stages of state and federal review, 24 had overall error rates of 52 percent or more.
• Two errors trigger the majority of reversals at the state post-conviction stage: "egregiously" incompetent defense lawyers who overlook evidence of actual innocence or evidence supporting a lesser sentence, and prosecutorial or police suppression of such evidence.
NOT A SUCCESS
In the end, the study concludes, few death sentences succeed, and it takes roughly 10 years for the courts to decide each one. From 1973 to 1995, states imposed 5,760 death sentences and carried out 313 executions -- 5.4 percent, or one in 19.
By that measure, the study says, capital punishment is not a success.
The 68 percent error rate -- which shows little fluctuation over 23 years -- is a "stunning number," says longtime death penalty scholar Ira Robbins, of American University Washington College of Law. It suggests that execution of the innocent -- the current public and legislative concern in the death penalty debate -- is not the only problem, he says.
"Whether or not the politicians agree, the courts are also reversing for serious constitutional errors," he says. "Legal innocence is still very important in the eyes of the courts."
The high error rates, Robbins says, show that "something is fundamentally wrong with the way the death penalty is administered in this country."
At least one proponent of the current system sees the error rate differently.
"The U.S. Supreme Court has said we're going to have super due process for death penalty cases," says law professor Paul Cassell, of the University of Utah, "and so it's not surprising that when you impose more process, you see more reversals. The death penalty system in America is the most accurate criminal sanction in the world."
Liebman, who also has represented capital defendants before the U.S. Supreme Court, counters, "If somebody came to you and said, 'I've got the perfect system. It's designed to make mistakes and catch them well,' I don't think most Americans want to pay for that kind of a system."
Beth Wilkinson, a partner in the Washington, D.C., office of Latham & Watkins and a former Oklahoma City bombing prosecutor, says, "What really troubles me is not just the 68 percent, but the 82 percent of people who, after error was discovered, received a sentence less than death. It shows me not only are there errors, but that most of these cases appear to not have been death penalty or death-eligible cases to begin with."
The study, "A Broken System: Error Rates in Capital Cases," began in 1991, after the chairman of the Senate Judiciary Committee asked for data on the frequency of habeas relief granted in capital cases. The study's release falls on the eve of a Senate Judiciary hearing on death penalty legislation, including DNA testing in alleged-innocence cases.
The findings appear to dispel some old and recent notions about capital appeals.
Because of releases from Illinois' death row and the Illinois governor's moratorium on executions, it's been suggested that Illinois has more than its share of death penalty errors. But its rate was 66 percent, slightly less than average.
Litigators have long believed that federal courts offer greater protection than state courts, whose elected judges are seen as vulnerable to political pressures. However, of the 2,370 death sentences reversed by state courts, 90 percent were overturned by elected judges.
That figure, Liebman cautions, doesn't mean that federal review is unnecessary. Federal courts still reversed 40 percent of the cases passed through by state courts.
Finally, states with the highest rates of death sentencing don't have correspondingly high execution rates. Louisiana, Virginia, Missouri and Texas have low death sentencing but high execution rates. Wyoming, Idaho, Nevada, Arizona, Oklahoma, Alabama and Mississippi have high sentencing rates but low execution rates. Finding the reasons will be part of the study's next phase, says Liebman.
The study gives the first broad picture of state courts in capital appeals, says law professor Franklin E. Zimring, of the University of California at Berkeley.
He says that he is struck by the "enormous, inexplicable and much scarier" state courts variance in error-finding.
"If you've been sentenced to death, your odds of error-finding in a state appellate court vary from 1 in 10 to 9 in 10, depending upon what state you're in," he says. In Virginia, he notes, 90 percent of all capital cases are error-free, according to its Supreme Court, but in North Carolina, only 39 percent are passed on as error-free.
Virginia is at variance not only among the death-sentencing states, but also within the 4th U.S. Circuit Court of Appeals. Its state court error rates on direct appeals range from Virginia's low of 10 percent to 61 percent in North Carolina.
The circuit itself has found reversible error in 15 percent of capital habeas cases -- the lowest error-detection rate among the nine circuits that conducted habeas review during the study period.
The other circuits and their error rates on habeas were the 6th (100 percent, two cases), 9th (62 percent), 11th (50 percent), 10th (47 percent), 7th (43 percent), 8th (33 percent), 5th (32 percent) and 3rd (29 percent).
Kent Scheidegger, of the conservative Criminal Justice Legal Foundation, in Sacramento, Calif., argues that high reversal rates result from "turmoil in the law" and not deficiencies in trial courts.
"Nobody can conduct a trial 'right' when 'right' isn't defined until years after the trial is over," he says. The Supreme Court, he notes, disagreed from year to year from the 1970s to the mid-1980s over what the Constitution requires. And, he adds, sometimes a reversal is due to an appellate court's previously reversing a valid judgment. "This is particularly true in the 9th Circuit," he says.
But Wilkinson and habeas scholar Larry Yackle, of Boston University School of Law, believe the study will withstand scrutiny and should affect the debate. "What Congress, and in large measure the Supreme Court, have done with the death penalty is to add procedural complexity that, I guess, is meant to speed things up, but in fact has slowed things down," says Yackle. "I don't think anything they have done in the last 20 or so years has done anything to seriously enhance our ability to avoid these errors."
Related Chart: State and Federal Circuit Habeas Appeals and Reversals