JULY 23, 1993









Mr. Chairman and Distinguished Member of the Committee, I am pleased to be here today to discuss the issue of the possibility of mistake in capital cases.

Mistakes in capital cases fall into two categories. On the one hand, it is possible that an innocent person might be executed. It is the specter of this kind of error that opponents of the death penalty seek to raise before this Committee. But it is also possible to make quite a different kind of mistake in capital cases. A guilty capital murderer might be spared the ultimate penalty only to kill other innocent persons. A fair assessment of the issue of mistake in capital cases leads inexorably to the conclusion that the risk to innocent life from failing to carry out capital sentences imposed under contemporary safeguards far outweighs the speculative and remote risk that an execution might be in error.

Expanding the opportunities to raise claims of innocence on federal habeas would substantially increase the risk of this kind of error -- an error I call "mistaken commutations" -- by unduly interfering with the imposition of death sentences throughout the country. Indeed, it is conceivable that allowing prisoners under capital sentences to raise unlimited claims of innocence would make it impossible for the states and the federal government to carry out any executions at all. If traditional habeas opportunities are extended to claims of factual innocence, a condemned prisoner would apparently be allowed to file repeated, frivolous applications raising claims of innocence, each of which could entitle the prisoner to a stay of execution.

In my testimony today, I will discuss four issues. I will first review the minuscule risk that an innocent person might be executed. Second, I will address the far deadlier problems that result when the death penalty is warranted but cannot be imposed due to failure to enact constitutionally-sound procedures. Third, I will review the current law regarding raising claims of innocence. Finally, I will identify various deficiencies in proposals to expand habeas remedies for claims of innocence. My testimony will be limited to these issues regarding mistakes in capital cases.

My background in this area is as follows. I am currently an Associate Professor of Law at the University of Utah College of Law. Form 1988 to 1991, I served as an Assistant United States Attorney in the Eastern District of Virginia, where I was responsible for prosecuting federal criminal cases. From 1986 to 1988, I served as an Associate Deputy Attorney General at the United States Department of Justice, handling various criminal justice issues including issues relating to capital punishment. In particular, in 1987, along with then-Assistant Attorney General William F. Weld, I prepared a Report to the Deputy Attorney General on Capital Punishment and the Sentencing Commission, a comprehensive review of death penalty topics, including the issue of the possibility of mistaken executions. In 1988, along with then-Assistant Attorney General Stephen J. Markman, I published int he Stanford Law Review a detailed rebuttal to an article by Professors Bedau and Raledet concerning allegedly erroneous executions. I have also served as a law clerk to then-Judge Antonin Scalia and Chief Justice Warren E. burger, writing memoranda on numerous criminal cases, including death penalty appeals.


Given the fallibility of human judgments, the possibility exists that the use of capital punishment may result in the execution of an innocent person. The Senate Judiciary Committee has previously found this risk to be "minimal," a view shared by numerous scholars. As Justice Powell has noted commenting on the numerous state capital cases that have come before the Supreme Court, the "unprecedented safeguards" already inherent in capital sentencing statutes "ensure a degree of care in the imposition of the sentence of death that can only be described as unique."

While the widely-held view has been that state death penalties are imposed with extraordinary care and accuracy, that position has recently been challenged. In 1987, two avowed opponents of capital punishment, Professors Hugo Adam Bedau and Michael L. Radelet, published the results of what they claimed to be "sustained and systematic" research over many years purporting to show that the use of capital punishment entails an intolerable risk of mistaken executions of defendants who are factually innocent of the crimes for which they were convicted. According to the authors, 350 persons have been wrongly convicted of capital or "potentially capital" crimes in the United States during this century; and twenty-three innocent persons have actually been executed. Professor Radelet repeated these claims in testimony before the Senate Judiciary Committee in 1989. Because the Bedau-Radelet article has been widely cited by death penalty abolitionists, it is important to consider it in some detail.

In 1988 after extensive research, then-Assistant Attorney General Stephen J. Markman and I published in the Stanford Law Review a detailed rebuttal to the assertions of Bedau and Radelet. We agreed with Bedau and Radelet that the execution of even one innocent person would be a tragedy. But we concluded that "not only is the Bedau-Radelet study severely flawed in critical respects, it wholly fails to demonstrate an unacceptable risk of executing the innocent. To the contrary, it confirms -- as convincingly as may be possible -- the view that the risk is too small to be a significant factor in the debate over the death penalty."

The Bedau-Radelet article suffers from a number of flaws. To begin with, it uses a peculiar definition of "potentially capital" cases. The definition includes some allegedly erroneous rape executions, even though rape is no longer a capital offense under contemporary Supreme Court decisions. At the same time, the article excludes treason, a quintessential capital offense, apparently so that Bedau and Radelet would not suffer criticism for declaring the innocence of Julius and Ethel Rosenberg, as they did in a draft of their article.

The authors also have included a great number of cases from the early part of this century, long before the adoption of the extensive contemporary system of safeguards in the death penalty's administration, which greatly skews their analysis. Bedau and Radelet are able to identify only a very few "miscarriages of justice" in the decade since the Supreme Court upheld the constitutionality of the death penalty. Out of approximately 50,000 murder convictions during the period from 1977 to 1986, the authors point to only five cases where, they claim, a death penalty was wrongly imposed, and in none of these cases was the sentence actually carried out. Even if one accepts their claim that all of these convictions were mistaken, the authors' accounts of these cases demonstrate that current post-conviction procedures work well in discovering and correcting errors. After all, in each of these cases the mistake was discovered.

Moreover, Bedau and Radelet cite but a single allegedly erroneous execution during the past twenty-five years -- that of James Adams. A dispassionate review of the facts of that case demonstrates, however, that Adams was unquestionably guilty, as I will discuss in a minute. Thus, Bedau and Radelet have made no persuasive showing that anyone has been wrongfully executed since new capital punishment procedures were instituted in the wake of the Supreme Court's decision in Furman v. Georgia. (All of the capital sentencing schemes throughout the country have extensive post-Furman procedural safeguards in them.) In short, what the authors have done is comparable to studying traffic deaths before the adoption of traffic signals.

The overwhelming problem with the Bedau-Radelet study is the largely objective nature of its methodology and therefore of its conclusions. The authors' standard of innocence is simply their belief that a majority of "neutral observers," given the evidence at the authors' disposal, would judge the defendant in question to be innocent. An examination of the James Adams case, the only alleged post-Furman case of an execution of an innocent person, demonstrates that far from acting as neutral observers, Bedau and Radelet have leaped to embrace as "innocent" even the guiltiest of defendants.

James Adams was convicted of killing then robbing a Florida rancher in 1974. Adams was executed in 1984. Bedau and Radelet claim that Adams was "innocent," but do not mention the following salient facts:

* Adams was arrested shortly after the murder with money stained with blood matching the victim's;


* Adams claimed that the money was stained because of a cut on his finger, but his blood did not match the blood on the money;


* Clothes belonging to Adams were found in the locked trunk of his car stained with blood matching the victim's;


* Eyeglasses belonging to the victim were also found in the locked trunk of Adams' car;


* Adams told the police when arrested that the clothing and eyeglasses found in his trunk were his, but at trial changed his story and denied owning any of the items.


* A witness, John Tompkins, saw Adams driving his car to and from the victim's house at the time of the murder;


* Another witness saw Adams' car parked at the victim's house at the time of the murder;


* A few hours after the murder, Adams took his brown car to an auto shop and asked that it be painted a different color.


* Adams' principal alibi witness contradicted him on the critical issue of his whereabouts at the time of the crime.


While ignoring all of this evidence, Bedau and Radelet offer the following to "prove" Adams' innocence:

* A witness who identified Adams' car leaving the scene of the crime was allegedly mad at Adams -- but Bedau and Radelet do not mention that three other witnesses also saw Adams at or near the scene of the crime;


* A voice that sounded like a woman's was heard at the time of the murder -- but the trial transcript reveals that this was the strangled voice of the victim pleading for mercy;


* A hair sample was found that did not match Adams' hair -- but Bedau and Radelet state falsely that it was found "clutched in the victim's hand" when in fact it was a remnant of a sweeping of the ambulance and could have come from any of a number of sources.


James Adams was a murderer and was justly convicted. As a result, even after "sustained and systematic" research, there is absolutely no credible evidence proving that an innocent person has been executed since Furman v. Georgia.

Bedau and Radelet's other alleged instances of "innocent" persons executed in earlier parts of this century are equally dubious. In our 1988 article, we reviewed all eleven cases of alleged executions of innocent persons in which appellate opinions set forth facts proved at trial in detail sufficient to permit a neutral observer to assess the validity of Bedau and Radelet's claims, including all of the cases since 1945. That review demonstrated beyond any reasonable doubt that all eleven of those defendants were guilty of capital murder. A full review of all of those cases would unduly expand the length of this article. I simply refer the reader to the relevant portion of our article for further details.

It may be important, however, to cite just a few of the scholarly abuses that we found in the Bedau-Radelet article. To prove the "innocence" of one defendant, Everett Appelgate who was executed for murdering his wife with rat poison in 1932, Bedau and Radelet cited two sources; those sources in fact actually believed that Appelgate was guilty. In another case, that of defendant Sie Dawson, the authors stated, falsely, that there were no eyewitnesses to the crime. In fact, there was an eyewitness -- the victim's four-year-old son, Donnie, who had been beaten and left to die at the scene of the crime. When found a day later, Donnie told his father, the police chief, and a family friend that Sie Dawson had committed the murder. Seventeen years later, when interviewed by the St. Petersburg Times about the details of the crime, he explained how Sie Dawson had brutally killed his mother with a hammer. As another example, Bedau and Radelet cite a book to prove generally the innocence of Charles Louis Tucker, executed in Massachusetts in 1906 for stabbing a young girl to death during a robbery; the book actually says that the Governor's rejection of Tucker's clemency petition was "conscientious and admirable."

Finally, my favorite example of Bedau and Radelet's research comes from my home state of Utah and involves one of their sources cited "generally" to prove that Joseph Hillstrom was innocent. That source was a book published by Wallace Stegner entitled Joe Hill: A Biographical Novel. The foreword explained that the book "is fiction, with fiction's prerogatives and none of history's limiting obligations. . . . Joe Hill as he appears here -- let me repeat it -- is an act of the imagination." While citing a work of fiction is bad enough, even more startling is the fact that the novel strongly suggests that its protagonist, Joe Hill, is in fact a guilty murderer! This is not surprising, since Wallace Stegner published two magazine articles in which he gave his view that the real-life Joseph Hillstrom was a killer.

The Bedau and Radelet study collected all claims of the execution of an innocent person through the summer of 1991. One case since then deserves brief comment. On May 18, 1992, Time magazine ran a cover story about Roger Keith Coleman entitled "This Man Might be Innocent; This Man is Due to Die." The article suggested that Coleman was innocent of the brutal rape and murder of his sister-in-law. Coleman, however, was plainly guilty. Indeed, Time concealed from its readers the most compelling evidence of Coleman's guilt -- DNA evidence developed by Coleman's own hand-picked DNA expert.

In 1982, when Coleman was tried, DNA testing was not sufficiently developed to be used in his case. In 1990, the state trial judge granted Coleman's request to have the state's physical evidence tested by Dr. Edward Black of California, whom Coleman's lawyers described as a "highly regarded expert" with "particular expertise" in DNA testing and "the interpretation of semen evidence in sexual assault cases." According to Dr. Blake -- who, remember, was selected and paid by Coleman's lawyers -- Coleman is a member of a class of only 2 percent of the entire general population who could have been the source of the sperm found in the victim's vagina. And when this evidence is combined with the independent fact that both Coleman and rapist/killer have type B blood, the percentage is reduced from 2 percent to 0.2 percent.

This newly-discovered evidence of Coleman's guilt comes on top of the trial evidence that two foreign pubic hairs found on the victim matched Coleman's and that blood of the same type as the victim's was found on Coleman's pants. A full recounting of the overwhelming evidence Coleman's guilt presented at trial is found in Virginia Supreme Court's opinion affirming the death sentence.

Coleman's attorney's claimed to have found newly-discovered evidence of his innocence. But all of the evidence was presented to a federal district court judge who, after careful examination, concluded that Coleman could not produce even a "colorable claim of innocence." Governor Wilder, who has commuted death sentences when questions of guilt have arisen, also carefully reviewed Coleman's claims and denied his request for executive clemency. Roger Coleman was a guilty murderer who deserved to be executed.

Based on a review of the cases offered in the Bedau-Radelet study and the more recent case of Roger Coleman, there is absolutely no credible evidence proving that an innocent person has been executed in roughly the last fifty years. I have testified to this effect before the Judiciary Committees of the both the House and Senate.

My testimony was recently challenged in an article by Professor David Cole of Georgetown, who asked: "Unless Cassell is divine, how can he know?" Cole's position boils down to little more than warmed over Humean skepticism about the possibility of determining the truth about such matters as the guilt or innocence of criminal defendants. But in the real world people do make judgments about such issues on almost a daily basis. Perhaps recognizing the weakness of his rhetorical ripepost, Cole goes on to argue:

For Cassell to substantiate his claim that no innocent person has been executed in 50 years, he would have to study every execution over that period and determine that, in fact, all those executed were guilty (assuming such a determination could be made).

But Cassell has not even tried to do that. All he has done is to provide the prosecution's response to those cases identified by others as executions of potentially innocent persons. Even if one accepted as true [this position], Cassell would not have demonstrated that no innocent person has been executed in fifty years. ... I believe the burden should rest on Cassell to prove his claim.


Cole is right in identifying the burden of proof as important in these matters. But it falls squarely on the shoulders of those who would proclaim the existence of martyrs, not those who defend the regular operation of the criminal justice system. Before any person is executed in this country, twelve members of a carefully selected jury have to decide -- beyond a reasonable doubt -- that a defendant is guilty. That is why, after all, a convicted defendant is presumed to be guilty and bears the burden of showing otherwise on appeal. In addition, appellate judges and the governor of the state must affirm the decision before any execution is carried out. Given these findings of guilt, the burden appropriately rests those who would challenge the existing state of affairs in the eyes of the law and claim that an executed person was in fact innocent. Even the dissenting justices in the Herrera case conceded this point, explaining that "[w]hen a defendant seeks to challenge the determination of guilt after he has been validly convicted and sentenced, it is fair to place on him the burden of proving his innocence, not just raising doubt about his guilt."

Given the presumption that those who are executed were in fact guilty, then-Assistant Attorney General Stephen Markman and I could appropriately limit our review of cases to those in which a claim has been made that an innocent person was executed. Professor and Bedau have exhaustively catalogued such cases. Their efforts in this regard extend back to 1964, when Professor Bedau attempted to collect such cases. He continued an informal effort through 1983, when he teamed up with another opponent of the death penalty, Professor Radelet, to collect such examples. From 1983 to 1987, they conducted "sustained and systematic" research to find every case that could of a wrongly executed person. Their labors included examining the New York Times index back to 1900 as well as the capital punishment holdings of the New York Public library and other significant collections. They also sent requests for information to 47 governors and to criminologists, defense attorneys, and others active in capital punishment around the country. They also combed the personal papers of those who had done previous systematic research on this subject. In a recent book, they extended their research through the summer of 1991. It seems unlikely that an acknowledged case of the execution of an innocent person would have escaped their notice, particularly one in recent years. I have also attempted to keep abreast of the academic and popular literature regarding allegedly innocent persons who have been executed. It is these sources of information, then, that I have used as the basis for my statement that no innocent person has been executed in this country since at least 1945. It is interesting that Professor Cole, while challenging my position, was unable to cite any case to the contrary.

The questionable examples in the Bedau-Radelet article make an important point about the debate over mistaken executions. It is easy for opponents of the death penalty to allege, despite a unanimous jury verdict, appellate court review, and denial of executive clemency, that an "innocent" person has been executed. Such an assertion costs nothing and will help abolitionists advance their cause. As this review demonstrates, such claims should be reviewed with a healthy dose of skepticism.

The paucity of examples of innocent defendants who have been executed provides compelling evidence that the risk of mistaken execution is virtually non-existent. If opponents of the death penalty are able to produce no better examples of mistaken executions than those put forward by Bedau and Radelet, then the overwhelming majority of Americans who support capital punishment can rest assured that the criminal justice system is doing an admirable, if not indeed perfect, job of preventing the execution of innocent defendants.




While modern-day examples of executed innocent defendants remain as rare as unicorns, it is much easier to find evidence that failure to execute justly convicted capital murderers would produce fatal mistakes, mistakes that I will designate as "mistaken commutations." Capital punishment saves innocent lives in three ways: through its incapacitative effect, its deterrent effect, and its role in establishing a system of just punishment. Failure to carry out properly imposed death sentences, after reasonable judicial and executive review, would thus be a mistake, a mistake with consequences no less lethal than a mistaken execution.

Let me emphasize that I am not urging that every first degree murderer be executed lest he kill again. Our present system of capital punishment limits the ultimate penalty to certain specifically-defined crimes and even then, permit the penalty of death only when the jury finds that the aggravating circumstances in the case outweigh all mitigating circumstances. The system further provides judicial review of capital cases. Finally, before capital sentences are carried out, the governor or other executive official will review the sentence to insure that it is a just one, a determination that undoubtedly considers the evidence of the condemned defendant's guilt. Once all of those decisionmakers have agreed that a death sentence is appropriate, innocent lives would be lost from failure to impose the sentence.


A. Incapacitation.

Capital sentences, when carried out, save innocent lives by permanently incapacitating murderers. Some persons who commit capital homicide will slay other innocent persons if given the opportunity to do so. The death penalty is the most effective means of preventing such killers from repeating their crimes. The next most serious penalty, life imprisonment without possibility of parole, prevents murderers from committing some crimes but does not prevent them from murdering in prison.

At least five federal prison officers have been killed since December 1982, and the inmates in at least three of the incidents were already serving life sentences for murder. For example, in the federal government's maximum security institution in Marion, Illinois, two experienced correctional officers were murdered in separate incidents on October 22, 1983. Officer Clutts died in an unprovoked, vicious assault when stabbed approximately 40 times with a homemade knife by inmate Thomas Silverstein. At the time, Silverstein was being supervised by three correctional officers. Silverstein had already murdered three inmates while in federal custody, for which he received three life sentences.

Officer Hoffman was murdered by prisoner Clay Fountain. Fountain managed to slip off his handcuffs and stab one of the three officers escorting him back to his cell. The other two officers rushed in. One of these officers was injured and the other, Officer Hoffman, was killed attempting to protect his fallen comrade. Following this unprovoked, brutal stabbing, inmate Fountain waved his arms in a victory expression as he walked down the cell ranges in front of other inmates. This inmate was serving a life sentence for the murder of a staff sergeant while in the United States Marines. He had repeatedly engaged in extremely violent acts, including the murders of inmates in 1979, 1981, and 1982. He was serving three life sentences at the time he murdered Officer Fountain.

Some killers are also paroled, only to kill again. For instance, Eddie Simon Wein was sentenced to death in Los Angeles Superior Court in 1957. Instead of being executed, he was released from prison in 1975 to live in West Los Angeles. Within months, he began to attack and kill women in the area. Fortunately for other potential victims, his apprehension was swift. He was convicted in 1976 of first degree murder on one woman, attempted murder of another, and numerous sexual offenses. The woman who was killed by Wein and the women who were scarred by him for life would not have been victims if Wein had been executed as originally decreed. Here the death penalty would have saved an innocent life.

Statistics prove that his is not just an isolated example. Out of a sample of 164 paroled Georgia murderers, eight committed subsequent murders within seven years of release. A study of twenty Oregon murderers released on parole in 1979 found that one (i.e., five percent) had committed a subsequent homicide within five years of release. Another study found that of 11,404 persons originally convicted of "willful homicide" and released during 1965 and 1974, 34 were returned to prison for commission of a subsequent criminal homicide during the first year alone. Of course, these figures reflect recidivism by murderers in general, not the presumptively more dangerous population of capital murderers.

While it is impossible to determine precisely how many innocent lives the execution of convicted murderers has saved, the available data suggest that the number is not insignificant. Of the roughly 52,000 state prison inmates serving time for murder in 1984, an estimated 810 had previously been convicted of murder and had killed 821 persons following those convictions. Execution each of these inmates following their initial murder conviction would have saved 821 innocent lives. Of course, since only a fraction of convicted murderers receive the death penalty, the number of innocent lives would be substantially smaller. Our data published in 1988 suggest a conservative estimate of at least 24 innocent lives saved just in the last few decades from the incapacitative effect of capital sentences, more than the total number of "innocent" defendants that Bedau and Radelet claim have been executed in this century.


B. Deterrence.

While the innocent lives saved through the incapacitative effect of capital punishment are important, the penalty also saves far more innocent lives through its general deterrent effect. Support for the deterrent effect of the death penalty comes from four sources: logic, anecdotal evidence, deterrence studies, and the structure of our criminal penalties.

Logic supports the conclusion that the death penalty is the most effective deterrent for some kinds of murders -- those that require reflection and forethought by persons of reasonable intelligence and unimpaired mental facilities. Many capital offenses are quintessential contemplative offenses. Murder for hire, treason and terrorist bombings all require extensive planning. It stands to reason that capital punishment deters such persons more than the next most serious penalty, life imprisonment without parole.

Anecdotal evidence in support of the deterrent value of capital sentences comes from examples of persons who have been deterred from murdering, or risking a murder, because of the death penalty. For instance, Justice McComb of the California Supreme Court collected from the files of the Los Angeles Police Department fourteen examples within a four-year period of defendants who, in explaining their refusal to take a life or carry a weapon, pointed to the presence of the death penalty. For instance, Louis Turck was arrested for robbery. He had used guns in prior robberies in other states but simulated a gun in the robbery in Los Angeles. He told investigating officers that, although he had been in California for only one month, he was wary of the state's death penalty. He used a simulated gun because: "I knew that if I used a real gun and that if I shot someone in a robbery, I might get the death penalty and go to the gas chamber." Similarly, Jack Colevris committed an armed robbery at a supermarket about a week after escaping from San Quentin. He was soon stopped by a motorcycle officer. As an escaped convict with two prior armed robbery convictions, Colevris knew he faced another long prison term. But he did not use the loaded revolver on the seat next to him because, he said, he preferred a possible life sentence to a death sentence.

Other evidence reveals that some criminals committed or attempted to commit homicides specifically because of the absence of a death penalty. For instance, according to the Attorney General of Kansas, one of the contributing factors leading to the reenactment in the 1930's of the death penalty in Kansas for first-degree murder was numerous deliberate murders committed in Kansas by criminals who had previously committed murders in states surrounding Kansas, where their punishment, if captured, could have been the death penalty. Such murders in Kansas were admittedly made solely for the purpose of securing a sentence to life imprisonment in Kansas if captured.

More recently, in March, 1973, four men entered a warehouse complex in Landover, Maryland, and took numerous hostages. Five of the hostages were shot, but they all miraculously survived. Eleven others were pistol-whipped. One of the victims, who had been shot in the throat, later testified that one of the robbers "told us he had a hand grenade and was going to blow us all up. He said it didn't matter to him who died, since the worst that would happen to him was that he would be taken care of the rest of his life" in prison. While no hand grenade was found, the fact that the robbers shot five people clearly indicates that they were quite willing to kill in the recognized absence of death penalty. Numerous other examples could be cited, but the fundamental point is that innocent lives are at risk from the absence of a death penalty.

Statistical studies support the proposition that capital sentences, like other criminal sanctions, have a deterrent effect. To be sure, some statistical surveys, often conducted by opponents of the death penalty, have found no such effect. A detailed review undertaken in 1987 by then-Assistant Attorney General William F. Weld and me found that few, if any, of these studies relied on rigorous methodologies or adequately controlled from many variables that affect the homicide rate in the jurisdictions under consideration. Moreover, it appears to be common ground in the scholarly literature on deterrence that a statistically valid study should account not only for the response of criminals to penalties imposed by an outside authority (the so-called "demand for crime") but also for an outside authority's response to changes in crime (the so-called "supply of crime"). The "supply-demand" econometric studies that have been done to date accord with our intuition and support the conclusion that the death penalty deters homicide.

One of the most recent substantial econometric studies was performed by Professor Stephen K. Layson of the University of North Carolina at Greensboro, who analyzed data for the United States from 1936 to 1977. Layson concluded that increases in the probability of execution reduced the homicide rate. Specifically, Layson found that, on average, each execution deterred approximately eighteen murders. Layson's study of the United States data is consistent with his earlier study concerning the deterrent effect of capital punishment in Canada and with important empirical work in this area by Isaac Ehrlich and other scholars. These econometric studies are buttressed by a growing body of scholarly literature demonstrating that punishment has a deterrent effect on crime in a wide variety of settings, including "cohort" studies using data on particular individuals rather than aggregate crime rates.

Indeed, the premise that more severe penalties deter crimes is fundamental to our criminal justice system. This Committee, for example, has often responded to the problem of crime by proposing new legislation that would increase the severity of criminal sanctions for certain criminal activities. Penalties for drug trafficking and firearms offenses immediately come to mind here. If those enhanced penalties were appropriate because they would help deter those offenses, death penalties for homicides are likewise appropriate because they will deter some would-be killers.

If Layson's empirical work (which is strongly supported by our intuition, anecdotal evidence, deterrence theory, and the structure of our criminal penalties) is correct, the death penalty has deterred roughly 125,000 murders in this country in this century. Put another way, based on Layson's study, more than one thousand innocent lives have been saved each year because of capital punishment. This figure dwarfs the twenty-three innocent persons Bedau and Radelet claim have been executed in the same time period. More important, it demonstrates rather starkly that under any realistic risk assessment, the chance of a mistaken execution is tiny compared to the risk from the "commutation" of all death sentences.


C. Just Punishment.

Through the imposition of just punishment, civilized society expresses its outrage and sense of revulsion toward those who, by contravening its laws, have not only inflicted injury upon discrete individuals, but also weakened the bonds that hold communities together. Certain crimes constitute such outrageous violation of human and moral values that they demand retribution. It was to control the natural human impulse to seek revenge and, more broadly, to give expression to deeply held views that some conduct deserves punishment, that criminal laws, administered by the state, were established. The rule of law does not eliminate feelings of outrage but does provide controlled channels for expressing such feelings. People can rely on society to sanction criminal conduct and to carry out deserved punishment. The law's acceptability and effectiveness as a substitute for self-help depends, however, on the degree to which society's members perceive the law as actually providing just punishment for particularly serious criminal offenses.

As the Senate Judiciary Committee has previously recognized, "[m]urder does not simply differ in magnitude from extortion or burglary or property destruction offenses; it differs in kind. Its punishment ought to also differ in kind. It must acknowledge the inviolability and dignity of innocent human life. It must, in short, be proportionate."

Determining what sanction is proportionate and, therefore, what constitutes just punishment for committing certain offenses is admittedly a subjective judgment. Nevertheless, when there is widespread public agreement that the death penalty is a just punishment for certain kinds of murder -- as there is in this country today -- and when a jury acting under standards meeting constitutional requirements determines that a particular person has killed another under circumstances for which the legislature has found that death is the appropriate penalty, the resulting judgment is no less "just" because its validity cannot be scientifically proven.

The death penalty's retributive function thus vindicates the fundamental moral principle that a criminal should receive his or her just deserts. Through the provision of just punishment, capital punishment affirms the sanctity of human life and thereby protects it. Walter Berns has traced this process, and his analysis is worth quoting:

The purpose of the criminal law is not merely to control behavior -- a tyrant can do that -- but also to promote respect for that which should be respected, especially the lives, the moral integrity, and even the property of others. In a country whose principles forbid it to preach, the criminal law is one of the few available institutions through which it can make a moral statement and, thereby, hope to promote this respect. To be successful, what it says -- and it makes this moral statement when it punishes -- must be appropriate to the offense and, therefore, to what has been offended. If human life is to be held in awe, the law forbidding the taking of it must be held in awe; and the only way it can made to be awful or awe inspiring is to entitle it to inflict the penalty of death.



The preceding sections have demonstrated two points: First, that no innocent person has been executed in recent history; and, second, the failure to impose death penalties can result -- and has resulted -- in the death of innocent persons. This section briefly reviews the current accommodation that has been reached between the need to review claims of innocence while at the same time allowing the effective imposition of death penalties.

The current capital sentencing scheme guards against the execution of innocent persons in at least seven ways.

First, the system imposes a vast array of due process protections to assure that no innocent person is convicted of a crime. The integrity of the criminal trial process is protected by rules governing the admissibility of evidence, the requirement that a defendant receive the effective assistance of counsel, the placement of the burden of proof on the prosecutor, the requirement of proof beyond a reasonable doubt, and the guarantee of a verdict by a unanimous jury on the issue of guilt or innocence. Because of these safeguards, it is "unlikely indeed that a defendant today could go to his death with knowledge of undiscovered trial error that might set him free."

Second, with respect to the role of federal courts in the process, federal courts can review the sufficiency of the evidence supporting a guilty verdict to ensure that the evidence was sufficient to persuade a rational trier of fact beyond a reasonable doubt of every element of the charged offense. Federal courts can review the sufficiency of the evidence on habeas corpus and can reverse a state conviction where the evidence was so lacking that no "rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt."

Third, the federal courts will allow any prisoner who can demonstrate his "actual innocence" to seek relief under federal habeas for a procedural error at his trial, even if the prisoner has forfeited his right to seek relief on habeas corpus due to a procedural default in the state courts or an abuse of the writ. Fourth, prosecutorial misconduct will result in reversal of a conviction. Both federal and state prosecutors are constitutionally required to disclose exculpatory evidence to a defendant and impeaching information regarding government witnesses. Failure to disclose such information that is material to a defendant's case requires a new trial. Moreover, a prosecutor may not knowingly present false testimony and, indeed, has a duty to correct testimony that she knows is false. Thus, if government misconduct results in the conviction of an innocent person, a remedy already exists.

Fifth, in the event that evidence comes to light after a trial that casts doubt on the accuracy of the trial verdict, a defendant may file a motion for a new trial. For example, in the federal system Rule 33 of the Federal Rules of Criminal Procedure allows a defendant to file a motion for a new trial based on newly discovered evidence within two years of trial. All fifty states and the District of Columbia authorize motions for new trials based on newly discovered evidence, with time limits ranging from ten days after judgment to no time limits at all.

In the event that exculpatory evidence is discovered after the time limits have expired, the sixth -- and perhaps most important -- safeguard against the execution of an innocent person is the possibility of clemency. As the Fifth Circuit has recognized, "Executive clemency is the last link in the chain." In the federal system, the Pardon Clause, Art. II, ' 2, cl. 1, authorizes the President to pardon a prisoner or to commute a capital sentence to any term of years, for any reason. Clemency has long been available in the states as well. All states that currently authorize the death penalty provide for the possibility of clemency, either by statute or by constitutional provision. It is even possible that executive clemency can come in the form of a motion by the prosecutor involved in the case. It seems likely that any prosecutors would drop charges in a case of genuine proof of a defendant's innocence.

It is clear that executive clemency has been frequently granted in capital cases. The Bedau and Radelet article, although flawed in the respects noted earlier, indicates that in 129 "potentially capital" cases a defendant was pardoned or received other executive clemency on grounds the authors identify as related to innocence. To cite some recent examples of executive clemency, in 1992, Governor Wilder of Virginia commuted Herbert R. Bassette, Jr.'s death sentence because "I cannot in good conscience erase the presence of a reasonable doubt and fail to employ the powers vested in me as governor to intervene." A year earlier, Governor Wilder commuted the death sentence of Joseph M. Giarratano, Jr., apparently because of doubts about guilt. Of course, these pardons do not mean that Bassette and Giarratano were in fact innocent of the crimes with which they were charged. To the contrary, very strong cases could be made that they were guilty. But the critical point is that executive clemency is available to death row prisoners who can demonstrate substantial claims of innocence.

It is sometimes argued that "politics" (however defined) sometimes prevents justified pardons. But whatever merit the argument might have in other contexts, when a clemency petition of an arguably innocent death row prisoner is involved it has none at all. Governor Wilder undoubtedly described the view of all governors when he explained, "You don't even think about politics when weighing death penalty cases." Moreover, "politics" has no sway when claims of innocence are involved. Not even a caricatured death penalty advocate would argue for the execution of innocent persons. Executive clemency based on sound claims of innocence is, if anything, good politics. As Governor Wilder explained in commuting Bassette's death sentence, he acted to "express the conscience of the commonwealth." Executive clemency is clearly a strong protection against the execution of an innocent person.

On top of all of these safeguards, the United States Supreme Court has recently recognized the possibility of even another, seventh safeguard against erroneous executions: review of claims of innocence on federal habeas. The traditional approach to claims of innocence raised on federal habeas was to consider only legal issues, not factual issues pertaining to guilt or innocence. As Justice Holmes explained, "[W]hat we have to deal with [on habeas review] is not the petitioners' innocence or guilt but solely the question of whether their constitutional rights have been preserved." Chief Justice Warren announced the governing standard in 1963:

Where newly discovered evidence is alleged in a habeas application, evidence which could not reasonably have been presented to the state trier of facts, the federal court must grant an evidentiary hearing. Of course, such evidence must bear upon the constitutionality of the applicant's detention; the existence merely of newly discovered evidence relevant to the guilt of a state prisoner is not a ground for relief on federal habeas corpus.


In recent years, the federal courts generally interpreted Chief Justice Warren's statement as precluding habeas review of claims of innocence, at least where the claim was based on newly discovered evidence.

In January, the Supreme Court cast doubt on the vitality of this limitation and suggested that claims of innocence could be reviewed on federal habeas in extraordinary circumstances. In Herrera v. Collins, the Court reviewed a claim raised by petitioner Herrera that he was actually innocent of the capital murder In various opinions, the Court denied his claim in a six-to-three decision. Of importance for present purposes, however, is that six Justices suggested that a sufficiently persuasive claim of actual innocence would entitle a petitioner to federal habeas relief. The opinion for the Court, written by Chief Justice Rehnquist, did not deny Herrera's contention that executing an innocent man would be subject to federal habeas review. Instead, the Court chose to "assume, for the sake of argument in deciding this cases, that in a capital case a truly persuasive demonstration of `actual innocence' made after trial would render the execution of a defendant unconstitutional." The three dissenters (Justices Blackmun, Stevens, and Souter) went further, and would allow a prisoner to raise a claim on habeas if he could establish that he "probably is innocent." Justice White concurring agreed that "I assume that a persuasive showing of `actual innocence' made after trial ... would render unconstitutional the execution of petitioner in this case." Justice O'Connor concurred, joined by Justice Kennedy, to note that she "cannot disagree with the fundamental legal principle that executing the innocent is inconsistent with the Constitution." Justice O'Connor went on to explain that the evidence "overwhelmingly" demonstrated that Herrera was guilty and that, therefore, she did not need to reach the issue of what circumstances would allow a petitioner to press a claim of innocence. In short, six justices seem to indicate that, despite the earlier history that habeas review was limited to claims of legal error, an innocent defendant could raise a claim of innocence in appropriate circumstances. Other commentators have read Herrera in exactly this fashion.

Apparently the only question remaining after Herrera is how strong a showing of innocence a petitioner would need to make to obtain relief. The pivotal concurring opinion of Justices O'Connor and Kennedy, whose votes were necessary for Chief Justice Rehnquist's opinion to command a majority of the Court, argued that federal proceedings and relief must be reserved for "extraordinarily high and truly persuasive demonstrations of actual innocence." The precise contours of that language remain unexplored because, as the concurring opinion observed, "If the Constitution's guarantees of fair procedure and the safeguards of clemency and pardon fulfill their historical mission, it may never require resolution at all."

In sum, the current capital sentencing scheme has an extraordinary array of safeguards to prevent the execution of an innocent person.





Several suggestions have recently been made to expand the opportunities for raising claims of innocence on federal habeas. Justice Blackmun argued in his dissenting opinion in Herrera that the Supreme Court should simply announce that the federal habeas statute allows a federal district court to entertain a claim by a defendant that he is "probably innocent." Academic commentators have advanced similar suggestions.

An example of legislation embodying these views is S. 221, which Senate Metzenbaum introduced in January. S. 221 would require a district court "to promptly stay" any pending execution whenever it received an "application" raising a claim that a prisoner under sentence of death is "probably innocent of the offense for which the death sentence was imposed." The stay would remain in effect until the district court's decision was affirmed by the court of appeals. The district court would grant "any appropriate writ or relief" if the applicant established "probable innocence" and that the evidence supporting the claim of innocence "could not have been discovered through the exercise of due diligence in time to be presented at trial." S. 221 would supersede all other provisions of law.

Proposal such as these are untimely, unnecessary, and unwise.

It is untimely to expand federal habeas review for claims of innocence because it now easier than in any time in history for a death row prisoner to raise such claims in federal court. The prevailing law in this country had long been that a prisoner could not raise a claim of innocence on federal habeas. Given the Supreme Court's recent recognition in Herrera of the possibility of bringing such claims on federal habeas, legislative action directed toward the same end seems ironically inopportune. The federal courts will now begin to define the breadth of the "actual innocence" claim created by Herrera. The federal courts have extensive experience in the litigation of habeas claims in general and capital claims in particular. The country's federal judges should be given the opportunity to develop the procedures and dimensions of actual innocence claims. If Congress is not satisfied with the result, there will be time enough to act later.

Expanding habeas review for claims of innocence is unnecessary. No doubt there are cases in which an apparently innocent person was convicted of a capital crime. Proposals like S. 221, however, do nothing to change the risk of convicting an innocent person. Instead, they simply create procedures designed to reduce the risk of executing an innocent person. As noted earlier, the present procedures work extraordinarily well in insuring the only guilty murderers are executed. Innocence is only rarely at issue. No innocent person has been executed in this country for roughly the last half century. Nor has it been explained how expanded habeas remedies would accelerate the review of claims of innocence. These proposals are thus completely superfluous.

To this point, it might be asked what is the harm in having excess protection against executing the innocent. This brings me to my final and most important point: expanding habeas opportunities to review claims of innocence is unwise. Proposals to allow expanded review focus single-mindedly on the risk of executing an innocent person and are blind to the larger dangers posed from failing to carry out executions in a timely fashion. In the language I used earlier, such proposals would dramatically increase the risk of "mistaken commutations" in their quest to avoid a peril that has never come to pass in recent memory. Like the commander at Pearl Harbor who grouped his planes to avoid the risk of sabotage while ignoring the larger risk of airborne attack, such proposals concentrate on the risk of executing an innocent person while remaining oblivious to the damage they would do to our country's capital sentencing system. The relative risks in this area are obvious: while no innocent person has been executed for roughly the last fifty years, each year more than twenty thousand of the nation's citizens will be the victims of murder. The death penalty is the most visible deterrent to such crimes. Small reductions in the deterrent value of capital punishment will clearly affect many innocent lives.

Expanding habeas review of claims of factual innocence could substantially reduce the deterrent effect of capital punishment by making it extraordinarily difficult to execute even murderers about whom guilt is not in doubt. For example, under S. 221 an automatic stay of execution is triggered whenever a prisoner files an application claiming innocence -- even where the district court finds the application to be frivolous, repetitive, abusive, or utterly without merit. To comply with the minimal requirements of the bill, the application need only be supported by "sworn affidavits" raising newly discovered evidence suggesting innocence. But such "evidence" is routine part of capital cases nowadays. As Justice O'Connor explained in rejecting the sworn affidavits submitted in Herrera, "Affidavits like these are not uncommon, especially in capital cases. They are an unfortunate although understandable occurrence. It seems that, when a prisoner's life is at stake, he often can find someone new to vouch for him. Experience has shown, however, that such affidavits are to be treated with a fair degree of skepticism."

California Attorney General Daniel E. Lungren explained how S. 221 would thwart the implementation of the death penalty in testimony presented to the Senate Judiciary Committee:

Under the bill, an eleventh hour claim of "probable innocence" would ensure entry of an automatic stay of execution, regardless of how weak the assertions may be, how many rounds of litigation have already transpired, or whether a similar claim of newly discovered evidence had already been rejected in state court. A federal judge -- often years after the original trial -- would then review the sworn affidavits or documented evidence and decide whether the petitioner is "probably innocent." If the judge concluded the petitioner was not "probably innocent," the petitioner could appeal this ruling, leading to a new round of litigation while the automatic stay remained in effect. If the trial judge of appellate panel decided the petitioner was "probably innocent," then some "appropriate writ or relief" would issue. ... Further, there is no limit to the number of "probably innocent" claims that a petitioner may file under the bill. All the petitioner needs to show is that the new evidence "could not have been discovered through the exercise of due diligence in time to be presented at trial." Consequently, single or multiple "probably innocence" claims could be filed five, ten or more years after the trial. Moreover, S. 221 encourages and rewards last-minute applications since there is no requirement that the newly discovered evidence be presented at the earliest opportunity that it may have become available.


A trumped up claim occurred in the recent California capital case involving murderer Robert Alton Harris. Harris filed a motion in the federal court for an evidentiary hearing based upon a newly filed declaration by inmate Joey Abshire attempting to cast doubt on Harris's guilt. After an evidentiary hearing, the federal district court held that "not only did Abshire lack credibility, but the court finds, without a doubt, that he lied to this court." The Ninth Circuit affirmed this finding. This Committee is no doubt familiar with the unnecessarily protracted litigation involved in bringing the Harris case to a close. California Attorney General Daniel Lungren testified before the Senate Judiciary Committee that "the Abshire declaration and subsequent evidentiary hearing were nothing more than fishing expeditions which caused unnecessary litigation and further delay ...." Of course, there is no likelihood that the risk of being held in contempt of court or prosecuted for perjury could possibly deter condemned inmates from filing frivolous or fraudulent claims of innocence.

As written, S. 221 would also specifically encourage delayed presentation of claims of innocence. The bill requires that a prisoner establish only that the alleged new evidence "could not have been discovered through the exercise of due diligence in time to be presented at trial." If a prisoner discovers evidence after trial, nothing in S. 221 requires him to present it in a timely manner. The affidavits filed in Herrera provide an example of how S. 221 would foster abuse. As Justice O'Connor explained, the affidavits were "suspect, produced as they were at the eleventh hour with no reasonable explanation for the nearly decade-long delay." Yet under S. 221, Herrera's affidavits would have been timely filed simply because the evidence arose after trial. S. 221 would thus create incentives for sandbagging.

While proponents of expanded habeas review might contend that these problems could be surmounted by a more carefully crafted bill than S. 221, in fact most of these problems are endemic to any system that allows claims of innocent to be readily raised on federal habeas. Justice O'Connor has warned that "[u]nless federal proceedings and relief -- if they are to be had at all -- are reserved for `extraordinarily high' and `truly persuasive demonstrations of actual innocence' that cannot be presented to state authorities, the federal courts will be deluged with frivolous claims of actual innocence." The problem is that once federal judicial scrutiny of claims of innocence are allowed years after the initial trial, prisoners under capital sentence can be expected to manufacture such claims with an understandable industriousness. As Attorney General Lungren has warned, "This second-guessing would be virtually assured in every case, as a convicted individual under sentence of death would have nothing to lose by pressing a claim of new evidence."

Repetitive challenges to a prisoner's capital sentence have already become the norm in capital cases, and such claims often are filed shortly before a prisoner is scheduled to be executed. Sometimes it even appears that last-minute filings are part of a deliberate tactic to stall executions. The problem of delay in capital cases has already been discussed extensively elsewhere. The average time between imposition of a death sentence and execution is now almost eight years and more than two thousand prisoners are on death row awaiting execution. Expanded review of claims of innocence would clearly add substantial delay to a system that already moves at a snail's pace.

Beyond the problems of repetitive and delayed claims, lenient standards of review for federal habeas claims of innocence would create a new and serious intrusion of federal power into areas of traditional state responsibility. As the Supreme Court recently explained, "It goes without saying that preventing and dealing with crime is much more the business of the States than it is of the Federal Government, and that we should not lightly construe the Constitution so as to intrude upon the administration of justice by the individual States." A federal habeas action, of course, allows a single federal judge to overturn a death sentence and release a convicted murderer despite a unanimous jury verdict of guilty, affirmance of the verdict by the state courts, and denial of clemency by the governor of the state. If the final and dispositive determination of the issue of an individual's guilt of a state offense is to be made by a single federal judge years or even decades later, there would seem to be little point in conducting the state trial in the first place. Such review of claims of innocence would relegate the entirety of the state criminal process, representing a substantial investment of resources, to the level of a mere preliminary proceeding.

Of course, there is no guarantee that the federal judge who second-guesses a jury's unanimous finding of guilt will make the correct decision. Indeed, there are substantial reasons to believe that such a belated guilt-innocence determination that is made on collateral review is, if anything, less likely to be accurate. As the Supreme Court has recognized, "the erosion of memory and dispersion of witnesses that occur with the passage of time prejudice the government and diminish the chances of a reliable criminal adjudication." Moreover, such after-the-fact hearings will inevitably focus on "newly discovered" evidence that was not presented at trial. Such evidence is almost invariably unreliable. If a defendant is truly innocent, he will be fully aware of the circumstances surrounding his innocence and can present them at trial. Circumstances that are presented only later are questionable. As the Supreme Court has concluded, "It is ... reasonable to presume that there is something suspect about a defense witness who is not identified until after the 11th hour has passed." But proposals to expand habeas review would appear to open the door to unreliable hearsay and other dubious evidence. The Federal Rules of Evidence are so skeptical of claims of criminal responsibility offered to exonerate a criminal defendant that they are inadmissible "unless corroborating circumstances clearly indicate the trustworthiness of the statement." But if "probably innocent" is the standard of review, federal judges would be required to routinely receive such evidence without suspicion. Finally, it is not clear how federal district courts will discharge the difficult task of weighing the "hot" newly-discovered evidence supposedly demonstrating innocence against the "cold" trial evidence proving guilt.

All of this suggests that expanded opportunities to raise claims of innocence would produce cases in which a single federal district judge would erroneously conclude that a guilty capital prisoner was "probably" innocent. The safeguards against this kind of mistake are minimal. Although the state could appeal the finding, the district judge's decision would be essentially final because it would appear to be a "factual" finding subject to only to a very deferential "clearly erroneous" review on appeal. The district court would then have to release the prisoner and order a new trial because that is the only "appropriate relief" for an innocent person. It might well be impossible for the state to reassemble its case a decade after the first trial and the in-fact guilty prisoner would be set free to prey again on society.

The risk of such "mistaken commutations" is far more substantial than the risk that an innocent person will somehow slip through the enormous body of safeguards in the present system against an erroneous execution. The murderers on death row are, obviously enough, an extraordinarily dangerous population. In some states, it is indeed necessary for the jury to find a risk of future dangerousness before imposing a capital sentence. Statistics confirm what common senses suggests about the threat to public safety posed by those on death row. More than ten percent of those under sentence of death have received two death sentences and more than four percent have received three or more. Apart from their capital convictions, more than 69 percent of death row inmates had prior felony convictions. Of the 2,356 death row prisoners in 1990, 417 were on parole at the time of their capital offense; 161 were on probation; 139 had charges pending, 64 were prison inmates, and 36 had escaped from prison. The erroneous release of even one of these prisoners would pose a great risk to the law abiding public.

The mistaken release of guilty murderers should be of far greater concern than the speculative and heretofore nonexistent risk of the mistaken execution of an innocent person.



An examination of proposals to expand review of claims of innocence leaves the (hopefully mistaken) impression that they were often drafted not solely to prevent the execution of an innocent person but to stop the execution of all prisoners under sentence of death, no matter how guilty. To be sure, reasonable people can disagree about the propriety and efficacy of the death penalty in this country. But as a policy matter, that debate has been resolved in favor of capital punishment. The Congress, at least 36 states, the Supreme Court, and the overwhelming majority of the American public all support the constitutionality and desirability of the death penalty. If these proposals are an effort to reopen the death penalty debate, one hopes that the subject will be approached in a more direct fashion.

* * * * *

Mr. Chairman, that concludes my prepared statement, and I would be happy to answer any questions.