6.2.1 Edwin Borchard’s Convicting the Innocent

Edwin Borchard, a comparative and international law scholar at Yale, was a pioneer in American law wrongful conviction scholarship (Leo, Reference Leo2005). He approached wrongful convictions not as a criminal justice scholar, but with “a political calculation that narrow legislation would be more likely to be enacted” (Zalman, Reference Zalman2020: 133). His draft compensation legislation was “intended to limit the relief to cases in which the justice of an award is obvious” (Borchard, Reference Borchard1932: 419). As discussed more fully in Chapter 2.4, he translated a more generous European approach that often compensated those who were acquitted after being imprisoned before their trial into draft legislation that required claimants to prove their innocence and that they were not at fault for their own wrongful conviction (Borchard, Reference Borchard1914; Roach, Reference Roach2024a).

Borchard was successful in his political ambition. President Franklin Roosevelt gave him a pen that signed a federal compensation law enacted in 1938 during the Depression (Zalman, Reference Zalman2020: 132). The result was what contemporary critics called a “puny statute” with compensation capped at $5,000 a year (Hale, Reference Hale1932). Even with increases in this cap, only two federal exonerees have received less than a million dollars in total since that time (Gutman, Reference Gutman2021: 238–239). The tail of compensation – an administrative and civil law matter – wagged the criminal justice dog of wrongful convictions.

Borchard’s Reference Borchard1932 book examining sixty-five different wrongful convictions set the template for much subsequent scholarship (Leo, Reference Leo2005). He stressed that he would only examine cases of “completely innocent people” (Borchard, Reference Borchard1932: xii). In rejecting a free standing right to claim innocence in 1993, the United States Supreme Court noted that executive clemency was provided in forty-seven of Borchard’s sixty-five cases (Herrera v. Collins 506 U.S. 390, 1993). Borchard took care to exclude any controversial case but said in private correspondence to Felix Frankfurter that while “the name Sacco-Vanzetti will not appear in the book” that it was “my humble contribution to preventing another such case” (Zalman, Reference Zalman2020: 142). He also excluded any rape case because there was no scientific way to determine the veracity of witnesses (Borchard, Reference Borchard1932: xiv). Six of the sixty-five cases he examined were extremely obvious because the murder victim was discovered alive (Reference Borchard1932: xviii; Warden, Reference Warden2013). In Chapter 9.2, it will be seen that three recent back-from-the-dead cases drove wrongful conviction reform in China. Borchard’s rationing of case was shrewd, but it only extended justice to some. Marvin Zalman (Reference Zalman2020: 141) has concluded that Borchard’s motivation was to counter innocence denials that continues to be “chiselled into the pages of American law reports by eminent judges such as Learned Hand, Sandra Day O’Connor and Antonin Scalia.”

Borchard recognized that in the majority of cases, the convicted persons “were poor and in many of the cases their defence was inadequate for that reason” (Borchard, Reference Borchard1932: xx). Although a number of his cases involved the conviction of Black people, he did not identify racism as a cause of wrongful convictions. In contrast, Borchard’s contemporary, the German lawyer Max Hirschberg (Reference Hirschberg1940: 40) appealed to Freudian psychology to argue that “political, racial, religious hatred” that “lies hidden in the subconscious” was a particularly insidious cause of wrongful convictions.

Many of Borchard’s findings remain relevant today. Misidentification by witnesses was the leading immediate cause in twenty-nine cases, but other cases involved lying witnesses and flawed and biased expert testimony. In at least thirteen cases, no crime was committed (Borchard, Reference Borchard1932: xiv–xv). This finding was significant, given that much subsequent wrongful discourse has wrongly assumed that all wrongful convictions involve the identification of a wrong perpetrator, allowing wrongful convictions to be seen as a failure of both justice and crime control (Roach, Reference Roach1999a).

Borchard also recognized that false confessions were not only the product of “the third degree” but of “persuasion or suggestion,” especially when the accused was of “inferior intelligence” (Reference Borchard1932: xviii). He found that police or prosecutors were at fault in most of the cases but also that they often subsequently demonstrated “commendable zeal in correcting the wrong” (Reference Borchard1932: xix). Many accused who were innocent did not testify for fear of their criminal records being used against them. Borchard’s early identification of the immediate causes of wrongful convictions underlines the message of Chapter 3: namely, that societies have not effectively addressed many of them for fear of compromising too many successful prosecutions.

Despite his focus on proven innocence, Borchard did not advocate innocence as a ground for appeal. Rather, he lamented that many American appellate courts would only review convictions for legal error and did not have the powers of the English Court of Criminal Appeal to overturn convictions because they were unreasonable or on any ground that there was a miscarriage of justice. He also praised the ability of New York appellate courts to overturn convictions “against the weight of the evidence” or because “justice requires a new trial” (Reference Borchard1932: xxiii). As will be seen, the concept of proven innocence that Borchard advocated for strategic reasons as a basis for compensation legislation has migrated into post-conviction relief in the United States. It has also been embraced by various volunteer organizations that have assisted the wrongfully convicted. In other words, the focus throughout the justice system has increasingly been on proven innocence not on the safety of convictions or whether they constitute a miscarriage of justice.

6.2.2 Jerome and Barbara Frank

In 1957, Jerome and Barbara Frank published a collection of cases where the accused was clearly innocent. Like Borchard, they avoided the more controversial cases of Sacco and Vanzetti or the Scottsboro teenagers, and many of the cases they discussed had been the subject of executive clemency and legislative compensation (Frank and Frank, Reference Frank and Frank1957: 29–30, 73, 129). Despite William O. Douglas’s preface which proclaimed that “an insistence on fair procedure is the great truth by which Jerome Frank lived” (Reference Frank and Frank1957: 12), the book’s main argument was not for fair trial rights but that “the conviction and imprisonment of innocent men occur too frequently to be ignored by any of us” (Reference Frank and Frank1957: 31). Indeed, the Franks argued against the “Upper Court myth” that wrongful convictions could be corrected on appeal. Drawing on Jerome Frank’s long experience as an appellate court judge, they explained “the upper court must pay no attention to the testimony of the defence witnesses, and take as true the testimony of the government witnesses, even if that court suspects that they lied or were mistaken” (Reference Frank and Frank1957: 33). Unlike Borchard, they did not even attempt to argue for more generous standards to reverse convictions on appeal.

The Franks’ focus on proven innocence was not for lack of sophistication. As befits Jerome Frank’s role as one of the pioneers of legal realism, the book recognized some of the complex psychological causes of wrongful convictions. In earlier work, Jerome Frank had argued “facts are guesses” and that lawyers often “appeal to the crudest emotions and prejudices of jurors” (Frank, Reference Frank1949: 14, 122). Juries “have to guess which witnesses have told the true story” and “unconscious prejudices” may result in error. There was also a recognition that “memory does not ‘mirror’ the past; memory re-creates the past” (Reference Frank and Frank1957: 208). They recognized confirmation bias or tunnel vision by observing that for humans “suspended judgment is painful… So we rush to a conclusion, then stubbornly cling to it, disregarding all facts contradicting that conclusion” (Reference Frank and Frank1957: 63–65).

The Franks, unlike Borchard, were aware that an innocent person could make rational decisions to plead guilty (Reference Frank and Frank1957: 114). They were also aware of the problem of lying and snitching by incentivized witnesses, with one informant admitting to framing over forty people (Reference Frank and Frank1957: 114). They focused on those who could be proven innocent in order to strengthen their reform agenda, particularly with respect to making prosecutors more concerned about preventing the blatant injustice of convicting the innocent. Prosecutors concerned with factual as opposed to legal guilt (Packer, Reference Packer1968) should also be concerned about the conviction of those who were clearly factually innocent. The Franks appealed to an “aroused public interest” (Reference Frank and Frank1957: 38) about the injustice of convicting and executing the clearly innocent. The Franks, like Borchard, were progressives (Zalman, Reference Zalman2020) though Jerome Frank had as a judge confirmed the death sentences of Julius and Ethel Rosenberg for espionage as not shocking the community conscience (United States v. Rosenberg 195 F.2d 583 1952). Nevertheless, in the punitive environment of American criminal justice, they focused on proven factual innocence and not broader understandings of miscarriages of justice.

6.2.3 The Court of Last Resort

The idea of charitable innocence projects in the United States has its origins in the 1950s when Erle Stanley Gardner, the author of the Perry Mason mysteries, recruited a medical doctor, a polygraph expert and a detective to help him uncover wrongful convictions. The group was formed after a magazine article Gardner had written convinced California Governor Earl Warren to commute a death sentence. The man was never released, however, because an alternative suspect was never found. Gardner recognized that the search “was made far too late” and “by private parties who did not have the facilities available to the police” (Gardner, Reference Gardner1952a: 14). Today, innocence organizations face the same challenges. Like the Franks, Gardner recognized that “if there had been an erroneous finding of fact the defendant has no remedy” and that appellate courts are “forever bound by the finding of the trial judge or of the jury” (Reference Gardner1952a: 121). This disregarded the limited ability of appeal courts in other countries, such as England, to review the reasonableness of factual findings.

The Court of Last Resort often appealed to state Governors who “are responsible to the people” and who were primarily concerned that their decisions on pardons could become “a political liability at election time” (Reference Gardner1952a: 22). It would only take on cases if the prisoner was “penniless” and serving a life sentence. This was a strict rationing of justice that reflected the charitable nature of the enterprise. Like many American innocence organizations today, it did not “represent the defendant. We would stand ready to publish the facts adverse to the defendant if and when they were discovered” (Gardner, Reference Gardner1952a: 6).

Gardner stressed that all wrongful convictions constituted a “double wrong” because the guilty went free, while the innocent were convicted (Reference Gardner1952a: 20). He argued “it is as great a miscarriage of justice when a guilty man is wrongfully acquitted as when an innocent man is wrongfully convicted” (Gardner, Reference Gardner1958–1959: 34). Gardner worried that increased awareness of wrongful convictions of the innocent could “start the pendulum swinging to the opposite extreme” and make juries “expect the prosecution to produce proof which amounts to a mathematical demonstration” (Gardner, Reference Gardner1952a: 181). These comments gave this first innocence project a populist and crime control edge. They ignored that Borchard (Reference Borchard1932: xiv–xv) had found wrongful convictions where no crime was committed, and there was no true perpetrator to find.

The intended audience for the Court of Last Resort’s work was “the reading public” (Gardner, Reference Gardner1958–1959: 74), Governors (Gardner, Reference Gardner1952a: 232), parole boards (Reference Gardner1952a: 114, 153) and prosecutors (Reference Gardner1952a: 180) who would respond to public opinion. Gardner looked to science, including lie detectors, as a means of improving policing. Although a former defence lawyer, he complained that defence lawyers were too focused on finding “legal loop holes.” He worried that his Perry Mason mysteries portrayed the police and prosecutors in an unfavourable light (Gardner, Reference Gardner1958–1959). The early American innocence movement in the 1990s and 2010s would similarly take pains not to appear anti-police or anti-prosecutor (Gould, Reference Gould and Zalman2014).

In 1952, Gardner welcomed 350 volunteers from the Texas bar to work with the Court of Last Resort. He saw this as part of an attempt to improve the “public relations” of the legal profession (Gardner, Reference Gardner1952b: 483). With the exception of one full-time investigator, all the professionals who constituted the Court of Last Resort donated their time: “if a defendant today should seek to have this body of men investigate his case and should try to pay them for their time, the cost would be prohibitive” (Gardner, Reference Gardner1952a: 13). Gardner spent $50,000 of his own money on the Court of Last Resort. Despite its success, the Court of Last Resort dissolved after Gardner’s retirement. He had advocated for a public body, one that he wrongly believed existed in England but one that would be created in North Carolina in 2006, to consider claims of factual innocence adding “it shouldn’t be too hard to weed out the wheat from the chaff” (Reference Gardner1952a: 33). Proven innocence for Gardner was a means of rationing justice and one that made sense for a charitable and voluntary organization but also one that could be transferred to a publicly funded institution that focused on post-conviction relief.

6.2.4 Hugo Bedau and Michael Radelet and the Innocence Sceptics and Deniers

None of the above-mentioned pioneering works directly focused on the death penalty or the risk of executing an innocent person. This recognized widespread support for the death penalty as well as the complex legalism of the American death penalty with such cases typically going through nine stages of post-conviction review in a “labyrinthine legal process” (Garland, Reference Garland2010: 44–45). The Supreme Court intervened in 1932 to prevent the execution of the wrongfully convicted Scottsboro teenagers and other Black men after “legal lynchings” (Reference Garland2010: 214–215), but these cases were not mentioned in the early wrongful conviction scholarship. This would not be the last time that innocence movements avoided the subject of anti-Black racism.

The American death penalty was revived by the local “politics of backlash” after the Supreme Court invalidated most death penalties in 1972, and thirty-five states enacted new death penalties (Reference Garland2010: 233). The 1976 resumption of executions meant that the death penalty could not be ignored. After the Supreme Court in a 5:4 decision dismissed a challenge to Georgia’s death penalty on the basis of anti-Black discrimination in its administration where white victims were involved (McCleskey v. Kemp, 481 U.S. 279 1987), the focus shifted to innocence claims. In that same year, the long-time abolitionist Hugo Adam Bedau with Michael Radelet published an article in the Stanford Law Review that outlined 350 cases in which those convicted of capital crimes “have later be found to be innocent,” including twenty-three cases where the convicted person had been executed (Bedau and Radelet, Reference Bedau and Radelet1987: 24, 71). Neither of the authors were lawyers, and they excluded cases based on lack of fault or due process or failures of affirmative defences in a way similar to the focus by Borchard, Gardner and the Franks on factual innocence. Just under 150 of their cases had received a pardon remedy from the executive or legislature (Reference Bedau and Radelet1987: 49).

Just as the Supreme Court’s 1972 decision ruling capital punishment unconstitutional (Furman v. Georgia 408 U.S. 238) produced a strong reaction, so too did the Bedau and Radelet study. Two lawyers then working in the Reagan administration’s Department of Justice, Stephen Markman and Paul Cassell, argued that the study did not establish an unacceptable error rate. The true error rate, they claimed, with over 7,000 executions, was only a third of 1 per cent (Markman and Cassell, Reference Markman and Cassell1988: 146). They argued that wrongful convictions in capital cases were a relic of the past because the only execution in the last quarter of the century cited by Bedau and Radelet was a case of a person who Markman and Cassell argued was factually guilty (Reference Markman and Cassell1988: 124). Reagan’s Attorney General, Edwin Meese, was even more sceptical about innocence claims and argued “you don’t have many suspects who are innocent of a crime. That’s contradictory. If a person is innocent of a crime, then he is not a suspect” (Norris, Reference Norris2017: 2).

Professor Cassell went on to become a judge, a leading crime victim advocate and a scholar who continues to argue that the wrongful conviction rate is grossly overstated by those associated with the innocence movement (Cassell, Reference Cassell2018). Another innocence sceptic was prosecutor Joshua Marquis, whose estimate of a 0.027 per cent wrongful conviction rate was endorsed by Justice Scalia (Kansas v. Marsh 548 U.S. 163, 198; Marquis, Reference Marquis2005). The low numerators of such claims can be explained in part by the difficulties of correcting wrongful convictions, while the high denominators reflect extraordinarily high rates of imprisonment in the United States. Professor Cassell has made the valid point that the DNA exonerations inflate contemporary error rates because they include convictions before DNA testing was available to clear people at an earlier stage of investigations (Cassell, Reference Cassell2018: 838). The advent of the American registry of exonerations, however, has revealed that DNA exonerations are a relatively small minority of all American exonerations. There are just over 600 cases involving DNA in the registry compared to over 3,600 exonerations as of the end of 2024.

Professor Cassell also continues to assume that cases resolved by guilty pleas are much less likely to be wrongful convictions, despite increased recognition by the American registry and through mass exonerations that many people who may well be innocent make decisions to plead guilty. In any event, he blames those who make false guilty pleas for their wrongful convictions by arguing that they commit perjury when they admit their guilt in court (Cassell, Reference Cassell2018: 844–845). Professor Cassell also argues that wrongful convictions should be considered in light of the risk of wrongful acquittals. Again, this neglects findings from the American registry that suggest that as many as 40 per cent of remedied wrongful convictions involve crimes that never happened. Reforms designed to prevent this type of wrongful convictions do not inevitably result in increased wrongful acquittals. His approach, which seems to value the harms of wrongful acquittals and wrongful acquittals equally, is supported by public opinion (Garrett and Mitchell, Reference Garrett and Mitchell2023). Nevertheless, it discounts the state’s more direct responsibility for wrongful convictions than for wrongful acquittals. Both the Bedeau/Radaelet and the Cassell/Marquis sides of the polarized debate focused on factual innocence and factual guilt and were driven by the United States’ moralistic and punitive penal culture (Packer, Reference Packer1968; Whitman, Reference Whitman2005).

6.2.5 Peter Neufeld and Barry Scheck’s Actual Innocence and the Innocence Project

Barry Scheck and Peter Neufeld published Actual Innocence in 2000, eight years after they had founded the Innocence Project at Cardozo Law School. Neufeld and Scheck were activists and public defenders who had started challenging DNA evidence, including its handling in the OJ Simpson case, but who then applied the new technology to overturn wrongful convictions. They stated that “not long ago, to claim that innocent person had been imprisoned was audacious, even risky, a proposition that was close to unprovable” (Scheck, Neufeld and Dwyer, Reference Scheck, Neufeld and Dwyer2000: xvi). Like Borchard, they targeted the innocence deniers. They conceded that the Bedau and Radalet study included some cases of “legal innocence, in which a person participated in a crime was charged incorrectly with murder. Others involved actual innocence” (Reference Scheck, Neufeld and Dwyer2000: 288). Neufeld recalled how before DNA there were often whispers about possible guilt even after people like Rubin “Hurricane” Carter had been exonerated and freed. He argued, “if you want to do policy and not just do exonerations … you want cases that are simply non-controversial, where everyone agrees – prosecutors, judges, the defence – that these people are stone-cold innocent” (Norris, Reference Norris2017: 56).

Scheck and Neufeld saved their concerns about the death penalty to one of the final chapters of Actual Innocence. They noted that since the Supreme Court in 1976 “put aside its reservations about the fairness, cruelty and reliability” of the death penalty that “for every eight executed, one innocent person is freed not only from death row but from incarceration” (Scheck, Neufeld and Dwyer, Reference Scheck, Neufeld and Dwyer2000: 282). They did not hide their politics or their disgust with 1996 legislation enacted with bipartisan support that required “federal courts to presume state courts are right about many things that state courts often are wrong about. Everyone agrees that is a terrible thing for an innocent person to be imprisoned. Far worse, though, would be for a politician to take a moderate line on crime” (Reference Scheck, Neufeld and Dwyer2000: 283). They noted that both candidates in the 2000 Presidential election were “unwavering supporters of capital punishment” who embraced “the DNA fallacy – that truth can be poured out of a test tube and make every hard decision easy” (Reference Scheck, Neufeld and Dwyer2000: 331–332). Scheck and Neufeld, self-described New York liberals, might have pursued a broader miscarriage of justice agenda in another time and in another place. But they recognized that a focus on “stone-cold” actual, factual and proven innocence was necessary, given bipartisan and highly punitive American political culture.

To their credit, Scheck and Neufeld were well aware of the limits of DNA as a means to establish wrongful convictions. They acknowledged that most crimes did not produce biological evidence that could be tested. They predicted that “in a few years, the era of DNA exonerations will come to an end” (Reference Scheck, Neufeld and Dwyer2000: 323). Reforms starting with improved procedures for taking eyewitness identification must be enacted quickly if many “innocent people accused of crimes in which the criminal did not ejaculate, spit, bleed, or shed tissue” (Reference Scheck, Neufeld and Dwyer2000: 323) were not to wrongfully remain in prison. Like Borchard, the Franks and the Court of Last Resort, they focused on the clearly innocent. Unlike them, they were able to apply the new technology of DNA to rape cases.

Actual Innocence and the syllabus of the course they taught at Cardozo Law School and freely shared focused on the immediate causes of wrongful convictions, such as mistaken eyewitness identification, lying witnesses, false confessions, inaccurate forensic testing, lack of disclosure and inadequate lawyering. Their focus on immediate causes was driven by a desire to explore the closing window of DNA exonerations and to propose many of the practical reforms examined in Chapter 3 such as the use of video-taped and double-blind photo identifications (Reference Scheck, Neufeld and Dwyer2000: 98–99), better disclosure from and oversight of forensic labs (Reference Scheck, Neufeld and Dwyer2000: 161–162) and screening of jailhouse informers (Reference Scheck, Neufeld and Dwyer2000: 203).

To their credit, Scheck and Neufeld, who as students in the 1960s helped the Black Panther Bobby Seale and led protests against the Vietnam War (Reference Scheck, Neufeld and Dwyer2000: 9), did not ignore the role of racism in wrongful convictions. They called out President H.W. Bush for his use of Willie Horton to defeat the Democrat nominee Michael Dukakis in the 1988 Presidential election (Reference Scheck, Neufeld and Dwyer2000: 233). They described how 51 per cent of DNA exonerations involved cross-racial crimes with 85 per cent of these cases involving Black accused compared to Justice Department estimates that only about 15 per cent of sexual assaults involved Black accused and white victims (Reference Scheck, Neufeld and Dwyer2000: 265). Many of these wrongful convictions followed a familiar script of all-white juries rejecting the evidence of Black alibi witnesses and accepting mistaken identifications by white witnesses (Reference Scheck, Neufeld and Dwyer2000: 271). Scheck and Neufeld were deeply aware and concerned about anti-Black racism. But they were realists. They started their searing and passionate chapter on race with a quote from the majority decision in the 1987 McCleskey v. Kemp 381 U.S. 279 case, which used the floodgate arguments that recognizing racial bias in the administration of the death penalty would lead to similar successful claims against the entire American criminal justice system. It was clear that they did not approve such arguments, but they knew that a conservative majority of the United States Supreme Court had ruled racial discrimination out of bounds. What they had and what they would run with both in courts and legislatures with incredible and surprising success was actual innocence.

6.2.6 Brandon Garrett’s Convicting the Innocent

The consistent American focus on proven innocence is underlined by Brandon Garrett using the same title in 2011, Convicting the Innocent, used by Borchard in 1932. Both books are landmarks. Before becoming the leading American wrongful conviction scholar, Professor Garrett worked with Scheck and Neufeld’s Innocence Project. Like them, he was well aware of the limits of DNA. For example, his book contrasted Dallas, Texas, where biological evidence was well preserved and produced nineteen exonerations with Houston, Texas, which destroyed biological evidence when storerooms became too full and only had six exonerations (Garrett, Reference Garrett2011a: 12).

Garrett’s study raised many troubling questions. The first 250 DNA exonerees he examined included fifteen cases in which the accused was convicted at trial despite having DNA testing that excluded them from evidence at the crime scene. In another fifteen cases, the person was also convicted in the face of other forensic evidence excluding them (Reference Garrett2011a: 100, 163). These cases suggested that attempts to use the science of DNA to produce an unquestionable consensus about “stone cold” innocence might not always succeed. It raised questions about prejudicial reasoning by prosecutors, judges and juries that caused them to deny or discount potentially exonerating evidence.

Professor Garrett reported eighty-four exonerations involving “black or Latino offenders and white victims,” while noting that “there is a long history of discrimination surrounding charges of rape involving white women and black men” (Reference Garrett2011a: 73). Overall, 155 of the first 250 DNA exonerations were Black, 74 were white and 20 were Hispanic (Reference Garrett2011a: 279). Only 6 per cent of the first 250 DNA exonerations had pled guilty, and most of those had made false confessions (Reference Garrett2011a: 150). Subsequent work by the National Registry of Wrongful Convictions, which is not limited to DNA exonerations, would reveal false guilty pleas as a much more pervasive problem. Garrett’s DNA data set did not deal with imagined crimes that did not occur. As of the end of 2024, the American registry would record that 40 per cent of the over 3,600 exonerations it recorded involved crimes that did not happen. As discussed in chapter 8.8, such wrongful convictions raise distinct issues not raised in DNA cases (Henry, Reference Henry2020; Roach, Reference Roach2023a).

Garrett’s book demonstrated the bipartisan and crime control appeal of DNA exonerations. In 45 per cent of the 250 exonerations, DNA identified the real perpetrator who had subsequently been convicted of fifty-six rapes and nineteen murders, while an innocent person was jailed for their original crimes. He also noted that half of Innocence Project clients had their guilt confirmed by DNA testing (Reference Garrett2011a: 233). Those who sought DNA testing and achieved such results would be eligible for increased punishment under many DNA testing statutes. As with the Court of Last Resort, the focus on factual innocence as determined by DNA had a crime control edge, even while it freed those lucky enough to be able to prove their factual innocence.

Garrett ended his book with a proposal for a series of reforms. He noted that when North Carolina’s Governor announced the creation of the North Carolina Innocence Inquiry Commission, he explained “as a state prosecutor for more than fifteen years, I know that law enforcement’s greatest nightmare is to have an innocent person in prison or death row” (Reference Garrett2011a: 243). Garrett praised the commission for focusing on innocence and avoiding being “tangled in complex procedural bars or harmless error rules” like “postconviction courts” (Reference Garrett2011a: 243). He also noted how New Jersey courts since 1999 required warnings about cross-racial identifications in a case involving an eventual and alas typical DNA exoneration after a white woman misidentified a Black man as the person who sexually assaulted her (Reference Garrett2011a: 247–248). Like Scheck and Neufeld, he was careful to note that DNA was only available in a minority of cases (Reference Garrett2011a: 265).

Professor Garrett has since championed the recognition of an international right to claim innocence (Garrett, Reference Garrett2017a; Garrett, Helfer and Huckerby, Reference Garrett, Helfer and Huckerby2021). As will be seen later in this chapter, such a right has still not been unequivocally recognized by the United States Supreme Court (Herrera v. Collins, 506 U.S. 390, 1993). In Chapter 11.4, I will argue that an international right to claim innocence would have regressive effects in many other democracies, including the United Kingdom, which as discussed in Chapter 5.2 attempts to provide post-conviction relief not only for those who can prove their innocence but also for those whose convictions are unsafe or involve a miscarriage of justice. Proven innocence represents a restrictive rationing of justice (Calabresi and Bobbitt, Reference Calabresi and Bobbitt1978; Nobles and Schiff, Reference Nobles and Schiff2000). It may be the best that can be achieved in extremely punitive societies such as the United States and China but it makes less sense in many other countries. Section 6.3 will examine how the focus on proven innocence has been popular with American legislatures.

6.3.1 Congress

Although it does not use the term “innocence,” the 1996 Anti-Terrorism and Effective Death Penalty Act Pub. Law No. 104–132 (AEDPA) enacted in response to the Oklahoma City bombing and concerns about delays in executions caused by death penalty litigation restricts access to habeas corpus through concepts very similar to proven innocence. Section 104 amending 28 USC 2254 presumes that factual findings in state criminal courts are correct unless the applicant rebuts the presumption by “clear and convincing evidence,” a standard often used in American law to require innocence to be proven. If the factual claim was not raised in the state proceedings, the habeas applicant must establish again by “clear and convincing evidence” that “but for constitutional error, no reasonable factfinder would have found the applicant guilty of the underlying offence.” This effectively required the applicant to show both a clear constitutional violation and innocence. The law also restricted the admission of new evidence of factual innocence more than in many adversarial systems by requiring that the new evidence could not have been obtained by the exercise of due diligence at trial. This often requires the wrongfully convicted to pay for their lawyer’s mistakes. Even Federal court review of state convictions based on legal errors was restricted in the 1996 law to clear errors. All of this deference to state courts reflected the unrealistic optimism that Harvard law professor Paul Bator demonstrated in 1963 when he argued against the Warren Court’s extension of federal habeas corpus because he could image a day when it was not necessary to be suspicious of about racism from southern justice systems (Bator, Reference Bator1963).

President Clinton supported the 1996 restrictions on federal habeas corpus, which had originally been championed in Republican Newt Gingrich’s “contract with America.” The law’s enactment was speeded up to coincide with the one-year anniversary of the Oklahoma City bombing. As Bryan Stevenson (Reference Stevenson2002: 701) noted, the law was enacted “in an atmosphere of anger and fear.” The law received bipartisan support despite warnings that its one-year statute of limitations could preclude new DNA evidence that would prove factual innocence. The law was a political piling on against prisoners who sought federal habeas corpus. Even before its enactment, liberal dissenters had characterized federal habeas corpus law as a “Byzantine morass of arbitrary, unnecessary, and unjustifiable impediments to the vindication of federal rights,” including requirements of proven innocence (Coleman v. Thompson, 501 U.S. 722, 759 (1991)).

Congress’s 1996 habeas restrictions have been widely criticized (Scheck, Neufeld and Dwyer, Reference Scheck, Neufeld and Dwyer2000: 282–283) but were not repealed under either President Obama or Biden. Garrett and Phillips (Reference Garrett and Kaitlin2022, 1752, 1761) have called for the law’s repeal on the basis that it produced “a procedural labyrinth that blocks potentially meritorious claims.” The 1996 legislation can be seen as a form of American extralegalism (Roach, Reference Roach2011: 163–169) because it is extremely complex and promotes litigation while at the same time making the extralegal end of the continued imprisonment and even execution of a wrongfully convicted person more likely.

In 2000, Senator Patrick Leahy, a Democrat from Vermont, introduced the Innocence Protection Act co-sponsored by a Republican from Oregon (Norris, Reference Norris2017: 87). Such bipartisan support was explained by Peter Neufeld, one of the founders of the Innocence Project, as important because “even the most law-and-order people felt it’s just terribly tragic when an innocent person goes to prison” (Norris, Reference Norris2017: 158). In introducing the act, Leahy commented that “too often in this Chamber, we find ourselves dividing along party or ideological lines.” He stressed that this act was “about what kind of society you want America to be in the twenty-first century” and not “about whether, in the abstract, you favour or disfavour the death penalty” (Leahy, Reference Leahy2001: 1115). One of the prices paid for a bipartisan approach to wrongful conviction reforms in the United States was avoiding discussing the death penalty, given Red/Blue polarization on the issue (Garland, Reference Garland2010).

Attempts to remove Congress’s 1996 restrictions on habeas corpus were removed before the Innocence Protection Act was enacted in 2004 as part of the Justice for All Act, which included increased statutory rights for crime victims. The focus on DNA was presented primarily as a means of crime control. Part of the law called the Debbie Smith Act, named after a rape survivor, contemplated almost $2 billion would be spent to increase DNA testing and assistance for crime victims (HR Rep 108–711). This dwarfed the sums that would be spent under the Innocence Protection Act. The Kirk Bloodsworth Act, named after the first DNA exoneree from death row, provided only $5 million over five years for DNA testing for convicts.

The DNA testing provisions for the wrongfully convicted also had a sharp punitive edge. They required applicants “to assert, under penalty of perjury, that the applicant is actually innocent” (s.412). There was a mandatory and additional minimum of three years’ imprisonment if this assertion proved to be false. In addition, unsuccessful applicants could be held in contempt of court, sacrifice good conduct credit in prison and have the results transmitted to their parole board. Law professor and innocence project lawyer Daniel Medwed (Reference Medwed2022: 157) has argued that such restrictions “make sense” because “they dissuade guilty defendants from taking a flyer on the chance that degraded evidence or flawed test might spring them from prison” (Reference Medwed2022). There were also requirements for the chain of custody of the evidence that contained DNA, but no federal legislation to require that such evidence be retained. The Act increased compensation for those wrongfully convicted under federal law who could prove their innocence. The increase was from $5,000 for each year of imprisonment under the original 1938 law as advocated by Borchard to $50,000 a year with $100,000 for each year that the proven innocence had spent on death row.

The Innocence Protection Act also attempted to legitimate and support the death penalty by providing for federal grants to states to improve both defence and prosecutorial representation in capital cases. No grants would go to states that had abolished the death penalty. This recognized the ability of death row exonerations to undermine public support and confidence in the death penalty. The grant for prosecutors was a concession to Republican senators including Judiciary Committee Chair Orrin Hatch. These grants were much more generous than those under the Kirk Bloodsworth Act providing for DNA testing. They authorized $75 million for each of the next five years.

The law was enacted by a 393 to 14 vote in the House of Representatives and by a voice vote in the Senate. President George W. Bush signed the bill into law just before the 2004 election, again underlining its bipartisan support. In 2004, no one could be against Justice for All. Robert Norris (Reference Norris2017, 88) has described the Innocence Protection Act “as one of the innocence movement’s greatest policy successes.” This is true, but a comparative and close reading of the act suggests that it contained many crime control features, and its innocence reforms were paired with a much more extensive and expensive array of measures for crime victims’ rights. That said, the Innocence Protection Act may well be the best that could be achieved in the punitive American environment.

Federal law reform has stalled since 2004. A Wrongful Convictions Tax Relief Act was included as part of the Protecting Americans Against Tax Hikes Act of 2015. As with the Innocence Protection Act of 2004, this law had bipartisan sponsors and support. It applied to both compensation under legislation and also compensation obtained through litigation. Amending the Internal Revenue Code does not engage competing crime control interests.

Any examination of criminal justice laws or practices that stops with examining federal laws in the United States is radically incomplete, given that the vast majority of prosecutions occur under state laws. What follows is an admittedly selective examination of two states that have been most active in wrongful conviction reforms while retaining the death penalty.

6.3.2 North Carolina

North Carolina has been one of the most active states with respect to wrongful conviction law reform. The impetus for such reform was an ad hoc bipartisan group including law enforcement, prosecutors, victims’ rights advocates and defence lawyers convened by Chief Justice Beverly Lake of the North Carolina Supreme Court. Lake was a conservative whose reputation was “burnished by his tendency to wear a pistol in the courthouse and an incident in which he ordered a defendant’s mouth to be shut with duct tape” (Medwed, Reference Medwed2022: 244). The group called themselves the North Carolina Actual Innocence Commission. They agreed not to discuss the death penalty. A moratorium on the death penalty was subsequently debated but defeated by the North Carolina legislature in 2003. North Carolina has executed 43 people since 1976, has exonerated 12 people from death row and has more than 120 people presently on its death row.

The mandate of the North Carolina Actual Innocence Commission was to “make recommendations which reduce or eliminate the possibility of the wrongful conviction of an innocent person” (Mumma, Reference Mumma2004: 650, 655). The Commission’s first recommendations asked police to voluntarily adopt best practices with respect to double-blind and sequential eyewitness identification, in response to the leading cause of wrongful convictions revealed by DNA. These recommendations were eventually codified by the North Carolina legislature in 2007. In 2004, North Carolina enacted new open-file disclosure requirements on prosecutors in response to a capital case where exculpatory evidence was not disclosed. In 2009, this was expanded to apply to all accused represented by counsel and not only those in capital cases (Pollard and Gammell, Reference Pollard and Gammell2024: 9). In four remedied wrongful convictions, however, the state had exculpatory evidence but did not disclose it to the accused because the accused did not, as required under North Carolina, ask for the disclosure (Reference Pollard and Gammell2024: 18).

As with the federal law, North Carolina’s DNA laws have a crime control edge that requires a “sworn affidavit of innocence.” In 2011, the state established a Forensic Science Advisory Board in response to a wrongful conviction based on forensic evidence (Pollard and Gammell, Reference Pollard and Gammell2024: 10). North Carolina laws also require electronic monitoring of custodial interrogations in felony criminal investigations, and since 2023, there will also be recording of interviews with jailhouse informants. Despite recording requirements, six of the eleven remedied wrongful convictions in North Carolina since 2006 have involved false confessions. In one 2011 case, the accused signed a confession despite reading difficulties, and the accused’s lawyer failed to request and the prosecutor failed to disclose exculpatory material including a ten-year old’s recantation of her claims of being sexually abused (Reference Pollard and Gammell2024: 13). In another case, an accused who did not speak English on advice of defence counsel pled guilty of second degree-murder in an impaired driving crash where the police wrongly identified him as the driver. North Carolina has been a leader in wrongful conviction reforms, but not necessarily in preventing wrongful convictions.

6.3.3 Texas

Texas is the surprising “gold standard in innocence reform” (Norris, Reference Norris2017: 135). This may be related to both Texas’s continued use of the death penalty and the close to 500 exonerations in that state. Since 1972, Texas has sentenced almost 1,200 people to death and has executed 587 people, with the Death Penalty Information Centre listing 18 people as exonerated from its death row. There are close to 180 people on Texas’s death row. Stephen Saloom of the Innocence Project has explained that “arguably, the reason that states like Texas, for example, has been so good on wrongful conviction reform is because those who want to be tough on crime also want to be seem like they’re being fair at the same time” (Norris, Reference Norris2017: 135). The many innocence reforms enacted by Texas have generally received all-party support. Following the precedent of the Kirk Bloodsworth Act for DNA testing that was included in Congress’s 2004 Innocence Protection Act, two of the main pieces of Texas reform legislation are named for high-profile exonerees. High-profile and sympathetic exonerees can compete with crime victims for political attention and empathy (Roach, Reference Roach1999a).

The case of Timothy Cole could well be discussed in Chapter 7 on racism and wrongful convictions. He was one of many Black men wrongfully convicted for raping a white woman based on honest but mistaken eyewitness identifications, as well as blood type and hair analysis that were later proved by DNA evidence to be erroneous. The victim had described the rapist as Black and as having smoked cigarettes throughout the attack. She identified Cole from a flawed photo line-up. Cole’s picture was the only polaroid taken by a detective who believed that Cole, then a twenty-four-year-old Army veteran and Texas Tech student, resembled a composite sketch. The jury rejected Cole’s alibi defence supported by his brother and friends, as well as his argument that he could not be the rapist because his asthma prevented him from smoking. Timothy Cole was sentenced to twenty-five years’ imprisonment. He died from an asthma attack in prison a decade before his 2009 exoneration. Four years before Cole’s death in prison, another man wrote to judges and prosecutors confessing to the rape. Timothy Cole always professed his innocence and was denied parole as a result. Texas Tech erected a statute in his honour with the caption “And Justice for All.”

In response to this tragic case, Texas created two advisory panels on wrongful convictions named after Timothy Cole. Like the North Carolina Actual Innocence Commission, the Texas panels represented various stakeholders in the justice system, including prosecutors, defence lawyers and judges. A 2016 report to Texas Governor Greg Abbott recommended recording of interrogations, increased prosecutorial disclosure, training for police on proper eyewitness identification procedure and no reliance on drug field tests. Unlike North Carolina, Texas does not have a publicly funded CCRC-type commission with powers to investigate and refer alleged wrongful convictions back to the courts.

The Texas Forensic Science Commission was created in 2005 to regulate forensic sciences. Its mandate has been expanded by legislation to provide for additional coverage and accreditation and licensing of various forensic science providers (Texas Criminal Procedure Code s.38.01). It can also take disciplinary action against forensic science providers. The nine-member commission must include one prosecutor, one defence lawyer and representatives of various universities involved in the delivery of forensic science. It has recommended that bite mark analysis not be used (Thompson and Casarez, Reference Thompson and Casarez2016: 721).

The Texas Commission encountered controversies over a report that found many deficiencies in arson evidence that had been used to convict and then execute Cameron Todd Willingham for the murder of his young children. Governor Rick Perry, who had refused to commute Willingham’s death sentence, removed some members from the commission. Subsequently, Governor Greg Abbott declared that the Commission did not have jurisdiction to examine whether there was misconduct in Willingham’s case because he was convicted and executed before the commission was created in 2005. Willingham’s last words proclaimed his innocence (Dioso-Villa, Reference Dioso-Villa2013; Gianelli, Reference Gianelli2011). The restrictions placed by Texas Governors on the investigation of his case lend support to Michael Radalet and Hugo Bedau’s (Reference Radalet and Bedau1998: 120) argument that a “defendant’s possible innocence has begun to seem almost like an argument against clemency” or investigations because of desires not to “embarrass the state officials” responsible for the conviction and the execution.

In 2009, Texas required corroboration of testimony given by jailhouse informants. It strengthened this regulation in 2017 by requiring full disclosure of any benefits provided to jailhouse informants. This legislation stops short of prohibiting the use of jailhouse informants, an issue discussed in Chapter 3.3. In the same year, Texas enacted a law requiring the full recording of interrogations in felony cases. Again, as discussed in Chapter 3, the efficacy of such reforms depends on the ability of defence lawyers to identify any hold-back information that the police may advertently or inadvertently provide to the suspect and that may lend authenticity to false confessions. In 2011 and again in 2017, legislation was enacted to require best practices in eyewitness identification. In addition, the details of all out-of-court identifications were required to be introduced to offset the ease in which witnesses can often identify the accused in court. In 2013, Texas created a new writ to allow for the entry of new science that, if considered at the original trial, would establish on a preponderance of evidence that an applicant would not have been convicted. In 2015, the writ was expanded to apply when experts changed their opinions. This was accompanied by increased funding and training for public defenders to use the new writ (Thompson and Casarez, Reference Thompson and Casarez2016: 727–729). Texas has increased the maximum amount for compensation for wrongful conviction to $80,000 a year. In 2013, Texas enacted a law requiring pre-trial DNA testing of available evidence in death penalty cases (Thomas, Reference Thomas2015: 1064–1065). Like the federal Innocence Protection Act, this reform demonstrates how innocence reforms can both increase accuracy in capital cases and help legitimate the use of the death penalty. Texas has a better record of wrongful conviction reforms than many more liberal states that do not use the death penalty.

Another well-known Texas exoneration that resulted in innocence reform legislation involved Michael Morton, who was exonerated of killing his wife in 2011 after serving twenty-five years in prison. There was a lack of full disclosure of evidence that would suggest that Morton’s wife was killed by an intruder in his house. In 2011, an undisclosed bandana near the crime scene was tested for DNA and was found to contain both the victim’s blood and the DNA of a man convicted of a similar murder in Texas in 1988. The prosecutor who withheld exculpatory evidence resigned his judgeship and was sentenced to ten days in prison for his failure to disclose the exculpatory evidence. This was a rare form of criminal accountability for those who caused a wrongful conviction, something that, as discussed in Chapter 9, has also occurred in China. The 2011 legislation, named after Michael Morton, imposed broad statutory disclosure obligations on prosecutors.

As in North Carolina, however, all of this reform has not prevented all wrongful convictions. Between 2010 and 2016, there were 13 exonerations in Texas of people wrongfully convicted of murder and 16 of child abuse and another 150 in drug cases, many related to people pleading guilty on the basis of field drug tests (Texas Exoneration Review Commission, 2016: 6). The National Registry lists twenty-five exonerations in Texas in 2024. Most are drug cases involving the Harris County drug lab, but three involved murder cases. The registry listed 491 Texas exonerations since 1989 as of the end of 2024. One case involved a Hispanic man convicted of a repeat drunk driving offence and sentenced to twenty-five years’ imprisonment and imprisoned for ten years before his conviction was overturned on the basis of police misconduct in manufacturing drunk driving crimes to increase revenue (National Registry, Revat Vara). A 2020 conviction of a Hispanic man for assaulting police was overturned on the basis of a lack of disclosure of the officer’s body cam footage (National Registry, Gonzalo Gonzalez). Texas is still trying to execute Robert Roberson, a man with autism and a grade 9 education, on the basis of discredited shaken baby syndrome (SBS) evidence used to convict him of murdering his two-year-old daughter, despite evidence that she died as a result of an infection (Findley et al., Reference Findley2023). Roberson has received some bipartisan support, but Texas Attorney General Ken Paxton and Governor Greg Abbott, both Trump supporters, are defending Roberson’s 2003 conviction and pursuing his execution.

Norris, Hicks and Mullenix (Reference Norris, Hicks and Mullenix2023) have found that the number of exonerations and the number of innocence organisations have a statistically significant effect on increasing wrongful conviction law reforms. At the same time, they find some evidence that the type of bipartisan support seen in the federal legislation and in North Carolina and Texas may be eroding. They conclude that states “whose public opinion trends liberal are more likely to adopt these reforms and to adopt them more quickly” (Reference Norris, Hicks and Mullenix2023: 55). The implications of these changes will be examined in the last part of this chapter.

6.4.1 The United States Supreme Court’s Hypocritical Approach to Executive Clemency

In 1993, the United States Supreme Court in an opinion written by Chief Justice Rehnquist rejected innocence as a free-standing ground for federal habeas corpus relief. Chief Justice Rehnquist justified his position by citing that the majority of cases in Borchard’s Reference Borchard1932 study received pardons and concluding: “history is replete with examples of wrongfully convicted persons who have been pardoned in the wake of after-discovered evidence establishing their innocence” (Herrera v. Collins, 506 U.S. 390 (1993)). The promise of executive clemency as a failsafe for Leon Herrera was false. Texas executed him in 1993 with its Board of Pardons and Parole, saying that it could not entertain his innocence claim. Herrera’s last statement proclaimed his innocence adding “something very wrong is occurring tonight” (Berger, Reference Berger2005: 145).

Only five years after invoking executive clemency as a failsafe that made a right to claim innocence unnecessary, Chief Justice Rehnquist dismissed a due process challenge to Ohio’s clemency process, stressing that clemency was not intended to enhance the accuracy of convictions but was an act of grace (Ohio Adult Parole Authority v. Woodard 523 U.S. 272, 284 (1998)). Judicial deference to executive discretion is an important aspect of American extralegalism (Trump v. The United States, 603 U.S. 593 2024). Only Justice Stevens in the 1998 case was prepared to contemplate judicial intervention if the clemency process was “infected by bribery, personal or political animosity, or the deliberate fabrication of false evidence” (Ohio Adult Parole Authority v. Woodard 523 U.S. 272, 290 (1998)). By 2010, American courts were describing executive clemency as “inherently discretionary, subject to the whim, or grace, of the decision-maker. It is, in other words, a form of relief to which a prisoner has no right” (McKithen v. Brown, 626 F.3d.143, 151 (2nd Cir.)

The British courts have been less deferential to executive clemency and more committed to the rule of law. In a 1993 decision not cited by the United States Supreme Court in their 1998 decision rejecting a due process claim with respect to clemency, a British court maintained that “If, for example, it was clear that the Home Secretary had refused to pardon someone solely on the grounds of their sex, race or religion, the courts would be expected to interfere and in our judgment would be entitled to do so” (The Queen v. Secretary of State ex parte Bentley 1993 EWHC Admin 2). The Judicial Committee of the Privy Council has applied duties of fairness to the exercise of executive clemency in the British Caribbean, including duties to delay executions pending petitions to supra-national bodies (Lewis v. Jamaica 2000 UKPC 35). American extralegalism becomes clearer when comparative analysis is employed. In the end, Chief Justice Rehnquist used the promise of executive clemency as a reason in 1993 to reject a right to claim innocence, and only five years later, he hypocritically concluded that the executive need not concern itself with the accuracy of convictions when making clemency decisions.

6.4.2 Innocence-Related Commutations

The Court’s reliance on clemency in Herrera v. Collins failed to note significant declines in the use of executive clemency since the death penalty was reformed in the 1970s. Pardons by US Presidents have declined with Jimmy Carter granting 534 pardons during his term as President with George W. Bush granting only 74 during his two terms and Donald Trump granting 144 during his first term (New York Times, 2024d). As of the end of 2024, the Death Penalty Information Centre listed 365 commutations of death sentences since 1976 (Death Penalty Information Centre, Commutations since 1976). Most of these were blanket commutations, including outgoing President Biden’s commutation of thirty-seven federal death sentences in 2024 to avoid the repeat of the Trump executions of 2020, which will be examined later in this chapter.

Most innocence-related commutations were issued before 2010. They are far outnumbered by commutations related to traditional mercy concerns about the offender. Despite this, Daniel Medwed (Reference Medwed2022: ch 10) has argued that clemency may be the best hope for the innocent in light of the increased hostility of more conservative American Federal courts to innocence claims. He called for the use of more diverse boards to make clemency recommendations (Medwed, Reference Medwed2022: 209). There are generally no special procedures for establishing innocence at the clemency stage. When innocence is mentioned, the standard of proof is often impossibly high with Georgia referring to “complete innocence”; New York referring to “overwhelming and convincing proof of innocence not available at the time of trial” and Texas requiring “either evidence of actual innocence from at least two trial officials; or the findings of fact and conclusions of law from the district judge in a state habeas action indicating actual innocence” (Cooper and Gough, Reference Cooper and Gough2014: 92–94). These standards are even higher than the clear and convincing evidence standard often required by courts. Even when clemency is used in this way, this affirms that proven innocence is a severe and populist rationing of justice.

Andrew Cuomo in his last days as New York Governor in 2021 commuted the 1998 murder sentence of Jon-Adrian (J. J.) Velazquez. Cuomo made no reference to whether Velazquez was innocent, despite concerns raised in the media since 2002 and celebrities championing his cause. Rather, he cited Velazquez’s good works in prison as a justification. After his release, Velazquez played a starring role in the movie Sing Sing, depicting how he and other prisoners formed a theatre group in prison. In 2022, President Joe Biden apologized to Velazquez for his wrongful conviction. It was only in 2024 that New York courts with the consent of the local District Attorney vacated Velazquez conviction on grounds of innocence. The reporter who originally raised concerns stated: “It’s a travesty that it took this long for a guy with obvious evidence of innocence and a ton of publicity, and it still took an army to get him exonerated” (New York Times, 2024a).

Executive clemency can play an important role in historical cases. In 1987, the Nebraska governor pardoned a man executed in 1887 for killing his work partner. A body had been misidentified, and the supposed victim had simply left for Mexico. In 2021, the Maryland Governor granted a post-humous pardon for John Snowden, a Black man hanged in 1919 for the rape and murder of a white woman. At the same time, the actual pardon declared the conviction and hanging a miscarriage of justice and stated that the case was too old to make a declaration of innocence. In 2021, Virginia’s Governor pardoned the Martinsville Seven, Black men who falsely confessed to raping a white woman under threats that they would be released to a lynching and who were executed in 1951. The pardon did not recognize their innocence, but rather that they had racially biased trials and punishment. It will be suggested in Chapter 7.8 that attempts to correct historical injustices may be one way to recognize the role that racism has played in many wrongful convictions.

6.4.3 Clemency for Miscarriages of Justice

Clemency is flexible and does not have to be limited to proven innocence. Governors have granted pardons to those convicted of possession of marijuana. A New Jersey governor who was not running for re-election created a clemency advisory panel with orders to expedite pardons for victims of domestic or sexual violence or sexual trafficking; those subject to excessive penalties; as well as those recommended by review by the Conviction Integrity Unit of the state’s Attorney General office. The panel was unveiled on June 19, 2024, with the Governor Phil Murphy stating: “As we celebrate Juneteenth and reflect on our nation’s ongoing journey toward racial justice for Black and Brown Americans, I am proud to sign this Executive Order to help address inequities and unfairness in our system of justice in New Jersey.” Just before Christmas 2024, he granted pardons to three women who were victims of domestic violence and were serving 60–30-year sentences for killing abusive partners. The women were represented by the American Civil Liberties Union, which formed a Clemency Project and not by an innocence organization (New York Times, 2024b). One possible future direction for innocence organisations would be to lend support for mass pardons or exonerations.

6.4.4 Political Polarization and Clemency

Recent experience with clemency has been affected by political polarization. The Republican Governor of Missouri Mike Parson, a former sheriff, has boasted that he has pardoned over 800 people. They include two white lawyers, Mark and Patricia McCloskey, who brandished their guns in fear of a Black Lives Matter protest and a white police officer Eric DeValkenaere convicted of manslaughter for shooting a Black man. Parson did not pardon Marcellus Williams, who was executed in 2024 despite widespread concerns about his innocence that will be discussed in greater depth at section 6.9.6. He also did not pardon three other Black men, Kevin Strickland, Lamar Johnson and Christopher Dunn, despite prosecutors and courts concluding that they were wrongfully convicted of murder and innocent (Shelly, Reference Shelly2025). Executive clemency can, as in New Jersey, be used as a vehicle to recognize the impact of anti-Black racism in the American legal system, but as in Missouri, it can also be used as a vehicle to express anti-Black racism.

6.4.5 Summary

In summary, executive clemency has sometimes provided a basis for the recognition of proven innocence, underlining from a legal process perspective the populist attractions of the concept. In a display of hypocrisy, the United Supreme Court has been quick to point to clemency as the ultimate fail-safe that makes it unnecessary to recognize a due process right to establish innocence in court while also refusing to ensure even minimal due process in the executive granting of clemency. Although the use of clemency has declined, the competing uses of it in 2024–2025 with respect to Biden’s commutation of thirty-seven federal death sentences and Trump’s pardons to over 1,500 people who supported him in the January 6, 2021, attempt to overturn his 2020 electoral defeat may make executive clemency a less credible mechanism to recognize innocence or perhaps other broader miscarriages of justice. In other words, executive clemency will be affected by political polarization and misinformation.

6.5.1 Free-Standing Claims of Innocence in Federal Courts

In 1993, the Supreme Court denied relief to a prisoner on Texas’s death row, claiming innocence and that his constitutional rights to due process would be denied by executing an innocent person. Chief Justice Rehnquist stressed that the Court had never recognized free-standing claims of innocence. Assuming that there was a constitutional right to claim innocence before execution, he stated:

He warned of the dangers of ordering new trials “simply because of a belief that in light of the petitioner’s new-found evidence a jury might not find him guilty at a second trial” (Reference Roach2011). The court stressed deference to the jury (Nobles and Schiff, Reference Nobles and Schiff2000) and deference to the states, including the death belt of former Confederate and Jim Crow states (Ogletree, Reference Ogletree2002).

Justice O’Connor with Justice Kennedy seemed more willing than Chief Justice Rehnquist to recognize a free-standing claim of innocence. Nevertheless, they similarly warned that unless the burden of proof on the applicant was “‘extraordinarily high’… the federal courts will be deluged with frivolous claims of actual innocence” (Herrera v. Collins 507 U.S. 390, 426). This recognized the high-volume nature of the American criminal justice system. They sought to leave open the question of whether the Constitution would prohibit the execution of an innocent person. Justice White was the only judge to clearly define a required burden of proof of innocence. It was the high standard “that no rational jury would convict when faced with newly discovered evidence and the entire record” (Herrera v. Collins 507 U.S. 390 429). The Court’s refusal to rule out a constitutional right to claim innocence or settle on the burden of proof spawned more litigation.

Justice Scalia with Justice Thomas would have preferred for the Court to rule that the Constitution did not forbid the execution of the innocent. They stressed that neither text nor tradition supported the right. They expressed confidence that the executive would grant relief should any persuasive claim of innocence ever be made.

Justice Blackmun with the concurrence of Justices Stevens and Souter dissented. They would have recognized a constitutional right to claim innocence based on the idea that executing an innocent person would be so shocking to the conscience that it would constitute cruel and unusual punishment under the Eighth Amendment. The dissenters cited in support the idea that the courts were already conducting a fact-based inquiry into actual or factual innocence with respect to allowing successive or defaulted federal habeas petitions. They ended on a chilling note that revealed the extralegalism of the majority’s approach when they wrote: “The execution of a person who can show that he is innocent comes perilously close to simple murder” (Herrera v. Collins 507 U.S. 390 446).

In 2009, the Supreme Court ordered an original writ of habeas corpus for the first time in fifty years. It remanded a capital case to a Federal District Court for the Southern District of Georgia with the instructions that the court “should receive testimony and make findings of fact as to whether evidence that could not have been obtained at the time of trial clearly establishes petitioner’s innocence” (In Re Davis, 130 S.Ct. 1 2009). This recognized a limited recognition of a free-standing claim of innocence but one that stressed, as in Herrera, a very high standard of proof. In a concurring judgment, Justice Stevens with Justice Ginsburg and Breyer seemed prepared to raise the burden higher and appeal to a false notion of certainty by stating: “imagine a petitioner in Davis’s situation who possesses new evidence conclusively and definitively proving, beyond any scintilla of doubt, that he is an innocent man” (In Re Davis, 130 S.Ct. 1, 2). Re Davis had the potential to generate what Cover and Aleinikoff had argued in 1977 was a genuine, if restrictive, focus on innocence where the courts would replace the executive that Chief Justice Rehnquist relied upon as a “safety valve” for the innocent. Alas, the District Court found that Davis had not established his innocence by clear and convincing evidence, and as will be discussed in Chapter 7.3, he was executed to international condemnation. In subsequent years, the Federal courts have failed to develop a habeas jurisprudence that takes innocence seriously (Litman, Reference Litman2018: 490–491).

Consistent with their 1993 Herrera opinion, Justice Scalia with Justice Thomas dissented. They stressed that Davis had been convicted of murdering a white police officer by a unanimous jury; had many post-conviction proceedings and been denied clemency. Quoting Paul Bator, they stated: “For much of our history, federal habeas review was not available even for those state convictions claimed to be in violation of clearly established federal law” (In Re Davis, 130 S.Ct. 1, 4). They concluded that the District Court in its search for new exonerating evidence was being sent on a “fool’s errand” because the 1996 Anti-Terrorism and Effective Death Penalty Act denied it the power to grant relief even if persuaded of Davis’s innocence. Their conclusion that Congress had stopped the Federal courts from preventing the judicial murder of an innocent person did not trouble them. They stated that an argument made by liberal judges that there might be an innocence exception not found in the text of the AEDPA did “not warrant a response” (In Re Davis, 130 S.Ct. 1: fn 2). This was the height of American extralegalism – the deployment of legalistic reasons to sanction the extrajudicial end of executing the innocent.

6.5.2 Free-Standing Claims of Innocence in State Courts

Only a year after the US Supreme Court rejected a free-standing innocence claim, Texas’s highest court recognized such a claim. The Texas Criminal Court of Appeal read Herrera as recognizing the right to claim innocence (State ex rel Holmes v. Texas Court of Appeal 885 S.W. 2d 389). Three years later, it extended this holding to non-death penalty cases with subsequent extensions to guilty plea cases. The initial threshold for a free-standing innocence claim was very high. It required that no rational jury would convict but was subsequently slightly lowered to require proof “by clear and convincing evidence that no reasonable juror would have convicted in light of the new evidence” (Ex parte Elizonda 947 S.W.2d 202, 1996). The different standards required to prove innocence allow for courts to ration justice (Calabresi and Bobbitt, Reference Calabresi and Bobbitt1978) and factor in the importance of the finality of convictions without full justification.

The Illinois Supreme Court, unlike the Texas Court of Appeals, held that the Supreme Court’s decision in Herrera precluded free-standing innocence claims under the Fourteenth Amendment’s due process clause, which applies to the state. Nevertheless, it ruled that the imprisonment of the innocent would violate both procedural and substantive due processes under Illinois’s constitution (People Washington 665 N.E.2d 1330, 1336, 1996). The legislature responded favourably to this decision by codifying free-standing innocence claims and imposing no statute of limitations on such claims. The bill’s sponsor argued, “we don’t want to have happen [again] what happened in [Illinois] where we had thirteen, and maybe even seventeen, people who were exonerated” from death row (Mundy, Reference Mundy2015: 22; 2003 Ill. Laws 4304). These alternative approaches demonstrate both the power of American federalism and the proven innocence paradigm.

6.5.3 Innocence Gateways to Federal Habeas Corpus Relief

Despite being reluctant to recognize a free-standing claim of innocence, American Federal courts have been much more enthusiastic in imposing innocence requirements on defaulted and successive attempts to seek habeas corpus relief in the Federal courts. Brandon Garrett has identified six different innocence standards used by American courts, noting that Justice Scalia had criticized them as “ineffable graduations of probability … beyond the ability of the judicial mind (or any mind) to grasp” (Garrett, Reference Garrett2011b: 1688ff). Different standards are used to disguise different judgments about the importance of the finality of convictions (Nobles and Schiff, Reference Nobles and Schiff2000). It is also an element of the United States’ complex extralegalism (Roach, Reference Roach2011: 163–169).

It is important to understand the Federal court’s development of innocence gateways in a historical perspective and as part of both legislative and judicial backlash to the Warren Court’s broad use of habeas corpus as a remedy to impose due process and equality values on all states, especially in the “death-belt” southern states. On the same day in 1963 that the Court released Gideon v. Wainwright extending the right to counsel to the states, the Warren Court delivered two judgments expanding the use of federal habeas corpus (Fay v. Noia 372 U.S. 391; Townsend v. Sain 372 U.S. 293). That very same year, Harvard Law professor Paul Bator warned that the Warren Court’s expansion of federal habeas corpus was undermining the need for efficiency and finality in applying the criminal sanction and the purposes of punishment. He argued that “an endless repetition of inquiry into the facts and the law” was a futile search for perfection and undermined the purposes of the criminal law and that “no court, can guarantee a correct result in an ultimate sense” (Bator, Reference Bator1963: 509–510). Bator argued that more faith should be placed in the traditional role of the elected executive as “a roving commission, usually free of technicalities and jurisdictional limitations, to seek out and right injustice” (Reference Bator1963: 525–526). Bator’s faith in executive clemency would influence Chief Justice Rehnquist’s rejection of a free-standing innocence claim in Herrera. His scepticism about ensuring that the innocent never be convicted also influenced Justice Scalia’s judgments.

Professor Bator got to the heart of the matter when he recognized that the Warren Court’s expansion of habeas corpus was driven in large part by its concerns for the rights of Black people and its suspicions that they may not be treated fairly by elected judges and juries in the Jim Crow south. That said, he then predicted that:

At a time when most schools remain segregated in the south and Bull Connor was using police dogs and fire hoses to attack peaceful civil rights marchers in Birmingham, Alabama, Professor Bator was, at best, astoundingly optimistic about the state of justice in the United States.

Bator’s concerns about finality and efficiency and cutting down on post-conviction relief were taken up by Judge Henry Friendly in 1970 at the start of the more conservative Burger Court era. Judge Friendly argued that the problem with expanded habeas corpus post-conviction relief in the Federal court was that the petitioner’s innocence had become irrelevant (Friendly, Reference Friendly1970). He noted that before the Warren Court, collateral attack on convictions was limited to a lack of jurisdiction. He criticized the use of federal habeas corpus as a means to vindicate the Warren’s Court 1961 decision imposing the Fourth Amendment exclusionary rule on the states and its 1966 Miranda decision extending the right to counsel to the interrogation room. In other words, he was defending what Herbert Packer (Reference Packer1968) identified as a crime control focus on factual guilt over the Warren Court’s due process focus on legal guilt. Finally, he championed a 1969 dissent by Justice Hugo Black that argued that he “would require that the convicted defendant raise the kind of constitutional claim that casts some shadow of a doubt on his guilt” (Kaufman v. United States 394 US 217, 242 (1969)). Judge Friendly was willing to impose a factual innocence requirement but, like Justice Black, one that at least accommodated the traditional reasonable doubt standard for proof of guilt (Friendly, Reference Friendly1970: 160). Friendly set some of the tone for bipartisan restrictions on federal habeas corpus relief that would finally be enacted by Congress and signed into law by President Bill Clinton in 1996 when he argued that “the system needs revision to prevent abuse by prisoners, a waste of the precious and limited resources available for the criminal process, and public disrespect for the judgments of criminal courts” (Friendly, Reference Friendly1970: 172).

In a 1986 case, the United States Supreme Court reversed a decision that had allowed federal habeas relief because of concerns about the use of a jailhouse informant. The Court in a decision that cited both the Bator and Friendly articles with approval reduced the scope of a general reference to the “ends of justice” made in a 1963 Warren Court decision to the need for a showing of factual innocence (Kuhlmann v. Wilson 477 U.S. 436, 454–456 1986). The liberal Justice Brennan and Justice Marshall dissented. They stated that the Court’s suggestion “that factual innocence is central to our habeas jurisprudence generally … was not supported by statutory language, legislative history, or our precedents” (Kuhlmann v. Wilson 477 U.S. 436, 462–463). They argued that habeas was concerned about rights and that Federal courts reviewing state convictions were not well equipped to make determinations of factual innocence. In 1989, Justice Stevens similarly rejected the idea that fundamental fairness could be reduced to “factual innocence.” He argued that “a touchstone of factual innocence would provide little guidance in certain important types of cases” (Teague v. Lane 489 US 288 at 321). Unfortunately, the history and wisdom of these dissents have been lost as the American innocence movement embraced factual innocence in an increasingly hostile and punitive environment that saw dramatic increases in prison populations (Garland, Reference Garland2001). Today, the beleaguered liberal wing of the Supreme Court defends innocence gateways without apparent recognition that they are fighting a battle in a war that has already been lost.

In 1991, the Court stressed that successive habeas petitions would, because of competing interests in the finality of conviction and deference to state processes, only be allowed in a “narrow class of cases” requiring a showing of factual innocence (McCleskey v. Zant 499 US 467. 495 (1991)). The same three liberal Justices again dissented. They argued that the Court was eroding the value of federal habeas corpus and was legislating a new standard that Congress had not articulated. This was Warren McCleskey’s second loss in the Supreme Court in four years. As will be discussed in Chapter 7.3, the same conservative majority of the Burger Court had in a 5:4 decision rejected his strong claim that Georgia’s death penalty was administered in a discriminatory manner that greatly increased the chance that he, as a Black man convicted of murdering a white person, would be executed (McCleskey v. Kemp, 481 U.S. 279 (1987)). The equality concerns that animated much of the Warren Court’s extension of habeas corpus were on the retreat.

In 1992, Chief Justice Rehnquist issued an opinion clarifying that courts should only consider successive habeas claims if the petitioner proved by clear and convincing evidence factual innocence (Sawyer v. Whitley 505 U.S. 333 1992). The Court noted that the “prototypical” example of “actual innocence” was “where the State had convicted the wrong person of the crime” and that the concept “was easy to grasp” (1992: 341). This appealed to the populist concept of factual or actual innocence. At the same time, it ignored the complexity of wrongful convictions including wrongful convictions for crimes that were not committed, something that Borchard (Reference Borchard1932: xiv–xv) had discovered. That said, proven innocence when coupled with crime control concerns that the real perpetrator had escaped had growing bipartisan appeal. Again, however, liberal judges dissented. They argued that the Court was applying too restrictive a concept of actual innocence, especially in a death penalty case (Sawyer v. Whitley 505 U.S. 333, 361).

By 1995, the Court clarified that the onerous clear and convincing evidence standard would not apply with respect to the innocence gateway for successive or defaulted habeas claims. Rather, the applicant would only have to establish probable innocence (Schlup v. Delo 513 U.S. 298). As the Court shifted, more liberal Justices became attracted to innocence-based exceptions. Justice Stevens commanded the majority of the Court and argued: “claims of actual innocence pose less of a threat to scarce judicial resources and to principles of finality and comity than do claims that focus solely on the erroneous imposition of the death penalty” adding that such claims “are rarely successful,” while “the individual interest in avoiding injustice is most compelling in the context of actual innocence” (Schlup v. Delo 513 U.S. 298, 323–324). Chief Justice Rehnquist, Justice Scalia, Justice Thomas and Justice Kennedy dissented. They would have insisted on the higher clear and convincing standard in all cases. The Court was greatly complicating the law by creating different standards for proof of innocence claims. It was placing a legalistic gloss on innocence while also stressing in all cases the importance of the finality of criminal convictions and deference to state processes.

In 1998, the Court extended the innocence gateway to cases where the accused had pled guilty over a dissent by Justice Scalia and Thomas. They pointed out, not unreasonably, that the standard would be impossible to meet without a trial record as well as more disruptive in a system where the vast majority of accused entered guilty pleas (Bousely v. US 523 U.S. 614 1998). Their dissent revealed that the majority may have rhetorically recognized and honoured the importance of innocence but through standards that almost all accused would fail to prove. In 1996, Justice Scalia commanded a majority in the Court that refused to create an innocence exception to a time limit on motions for judges to acquit people after the jury had rendered the verdict. He stated he refused “to fashion a due process right out of thin air” (Carlisle v. US 517 U.S. 416 at 430).

In a 2006 case, the Supreme Court found that an innocence gateway had been satisfied in a case where DNA evidence became available after a capital murder conviction (House v. Bell 547 U.S. 518). Innocence gateways did not require proof of innocence to an absolute certainty. At the same time, the case was an obvious injustice given that the DNA analysis excluded the person convicted, and there was a confession by the real perpetrator. Nevertheless, Justice Scalia with Chief Justice Roberts and Justice Thomas dissented. They argued that the convicted person had failed to establish that it was probable that no reasonable juror would convict after considering the new DNA evidence. The sort of innocence scepticism started in academic debates (Markman and Cassell, Reference Markman and Cassell1988; Marquis, Reference Marquis2005) was also influencing the more conservative members of the court.

In 2013, the Supreme Court in a 5:4 decision narrowly upheld the actual innocence gateway as a way around a one-year statute of limitations imposed in the 1996 legislation restricting habeas corpus. Liberal judges, who in the late 1980s and early 1990s resisted actual innocence as too restrictive a ground for allowing successive or defaulted claims, now cheerfully accepted actual innocence. Justice Ginsburg for the Court interpreted the 1996 act as modifying but not eliminating actual innocence claims by requiring them to be proved on the higher standard of clear and convincing evidence (McQuiggin v. Perkins 569 U.S. 383, 386, 396). She noted that the success of innocence gateways would be “rare.” Because of this, they achieved a balance between “the individual interest in justice that arises in the extraordinary case” and “finality, comity and conservation of scarce judicial resources.” In short, proven innocence was a means to ration justice to “rare” cases.

Justice Scalia dissented with Chief Justice Roberts and Justices Thomas and Alito. They argued that the court had no power to create an exception from the one-year limitation period in the 1996 law. Like Professor Bator in 1963, Justice Scalia argued that the majority was inspired by an unrealistic “vision of perfect justice” and that it would require Federal courts to wade “into the murky merits of the petitioner’s innocence claim.” He raised the spectre that the over 12,000 habeas petitions made to Federal court each year would increase even more.

The narrow exception of proven actual innocence is becoming a dissenting position as the Trump appointees move the Court in a more authoritarian direction. In a 2022 case, Justice Sotomayor complained in dissent that the Court’s decision not to allow an actual innocence exception forced the petitioner to pay for his lawyer’s mistake (Shinn v. Ramirez 596 US 366, 406 (2022)). Justice Thomas for the majority stressed both the state’s interest in finality and the brutal facts of the actual case.

In another 2022 case, Justice Gorsuch for the Court seemed to regret that the Federal courts ever ventured beyond review for jurisdictional error in applying federal habeas corpus. He quoted Paul Bator’s Reference Bator1963 article and suggested that a 1953 habeas case that involved the equality rights of a Black accused (Brown v. Allen 344 U.S. 443) was the start of a review of state court proceedings that led to an “an exploding caseload of habeas petitions from state prisoners.” He applied statutory restrictions in the 1996 legislation to uphold a decision that a first-degree murder trial conducted while the accused was in handcuffs, waist chains and ankles bracelet throughout the trial was not an unreasonable application of established federal law. The idea that Federal courts could only intervene for violations of clearly established federal law embraced a form of extralegalism that suggested that even weak claims of legality verging on the pretextual would be sufficient to sustain convictions and defeat damages or any other accountability measures when officials had violated the law (Brown v. Davenport 596 U.S. 118 2022; Roach, Reference Roach2011: 436–438). Justice Kagan in her dissent opposed the majority’s reading of history and pointed out that long before the 1953 case, Federal courts had used habeas corpus to vindicate constitutional rights.

In Jones v. Hendrix 599 U.S. 465 (2023), the majority of the Court rejected an attempt by the Biden administration to extend innocence gateway exceptions, noting that even the original exception was “bold” (Jones v. Hendrix 599 U.S. 465, 491). Justice Jackson in her dissent quoted from the legislative history of the 1996 legislation as evidence of bipartisan support for factual innocence exceptions. She quoted Republican Senator Orrin Hatch stating in 1995 that the law restricted federal habeas “unless, of course, they can truly come up with evidence of innocence that could not have been presented at trial. There we allow successive petitions.” Then-Senator Biden similarly explained that the goal of the 1996 legislation was “essentially giving one bite out of the apple to drastically reduce the ability to have successive petitions unless there is some egregious action that is learned about after the petition is filed, the first petition” (Jones v. Hendrix 599 U.S. 465, 508).

Factual innocence gateways in federal jurisprudence were not a progressive development. In the 1980s and early 1990s, they were resisted by liberal judges who were still fighting to retain some of the Warren’s Court broad expansion of federal habeas corpus as a means of ensuring that all states respected due process and equality. Restrictive innocence gateways may shrink in the future, given the composition of the United States Supreme Court. The few remaining liberal judges are attempting to hang on to such exceptions, but there are questions about whether such battles are worth it, given the difficulty of proving innocence, especially as DNA exonerations decline with increased investigative use of the technology.

6.5.4 Does the United States Supreme Court Care about Innocence?

Brandon Garrett’s research on the first 250 DNA exonerations found thirty-eight cases that were litigated before eventual exoneration all the way to the United States Supreme Court. In each and every case, the Court denied the wrongfully convicted person relief (Garrett, Reference Garrett2011a: 196). He added that no petitioner was able to make out a free-standing innocence claim and the six persons who tried all failed (Reference Garrett2011a: 203). Finally, he noted that only 3 per cent of the first 250 DNA exonerations “received DNA testing and a vacatur through federal habeas corpus – another indication of the small role played by federal courts” (Reference Garrett2011a: 230). Daniel Medwed (Reference Medwed2022: 128) has also concluded that the “Great Writ” “doesn’t deserve is nickname” because “its focus on constitutional and jurisdictional defects overlooks the fact-based, not law-based, nature of most innocence claims.” The terms “factual” or “actual innocence” have been mentioned in over fifty United States Supreme Court decisions. The conclusions of leading wrongful conviction scholars Garrett and Medwed, however, suggest that the judicial rhetoric of honouring innocence may be a form of false advertising and legitmation.

In 1977, Robert Cover and Alexander Aleinikoff conceded that even as the United States Supreme Court was cutting back on the Warren Court’s extension of federal habeas corpus in the name of due process and equality, it was possible to imagine that the court could create a principled jurisprudence around its newly asserted concern about innocence. At the same time, they expressed scepticism that the Court was heading in this direction. They believed that the Court was motivated more by concerns about efficiency and finality in the criminal process and deference to the states. The Court’s entire innocence jurisprudence, including its rejection of a free-standing claim of innocence and its restrictive jurisprudence on innocence gateways, confirms Professor Cover and Aleinikoff’s prediction made close to half a century ago.

The Supreme Court’s lack of genuine concern about innocence is also evident beyond its complex and legalistic habeas corpus jurisprudence. In 2009, the Court rejected arguments that access to post-conviction DNA was required by due process during the height of its professed concerns about actual innocence. It stressed the importance of deferring to the states and to the finality of convictions (Dist. Attorney’s Office for the Third Judicial Dist. v. Osborne, 557 U.S. 52, 62 (2009)).

DNA testing is only possible if evidence is retained. In 1988, the Court held that due process was only violated when the police acted in bad faith in not retaining semen and other biological evidence. Three liberal judges dissented and expressed concerns about the safety of the conviction for kidnapping and sexual assault of a child, noting that there was no other physical evidence linking the accused to the horrific crime (Arizona v. Youngblood, 488 U.S. 51, 71–72 (1988)). This was born out as other DNA evidence was eventually found and exonerated the accused twelve years later and convicted the true perpetrator (Medwed, Reference Medwed2022). Some states require a higher standard than bad faith with respect to the retention of evidence (Commonwealth v. Henderson 411 Mass. 309 (1991)).

The Supreme Court has not been sensitive to the dangers of false confessions. In 1986, it ruled that due process was not violated by the admission of a confession by an accused who was severely mentally ill even though the confession was the only evidence linking the accused to the crime. (Colorado v. Connelly, 479 U.S. 157, 170–71 1986). In 2018, it refused to hear Brendan Dassey’s claims that he, as a cognitively challenged teenager without a lawyer or parent present, had falsely confessed to be involved in a murder committed by his uncle, Steven Avery. The 7th Circuit had narrowly upheld Dassey’s conviction by stressing that Federal courts must defer to factual and legal findings by state courts. Strong dissents concluded that the Wisconsin court decisions were not entitled to such deference because of their brief conclusory reasons and that Dassey’s confession had all the hallmarks of a false confession (Dassey v. Dittman 877 F.3d 297, 320 (7th Cir) cert denied United States Supreme Court, June 25, 2018).

In 2012, the Supreme Court refused to increase protections against mistaken eyewitness identification in the face of strong evidence that it was the leading cause of wrongful convictions. Justice Ginsburg concluded that the court would only screen identifications for reliability “to deter police from rigging identification procedures, for example, at a lineup, showup, or photograph array” (Perry v. New Hampshire, 565 U.S. 228, 232–233 2012). Like its prior decisions on retention of evidence and confessions from vulnerable persons, the Court would only intervene if there was clear police misconduct that should be deterred.

Another case demonstrating the Supreme Court’s lack of genuine interest in protecting innocence involved a Black grandmother convicted on the basis of SBS of murdering her 7.5-week-old grandson. The Court concluded: “the State’s experts, whom the jury was entitled to believe, opined that the physical evidence was consistent with, and best explained by, death from sudden tearing of the brainstem caused by shaking” (Cavazos v. Smith 565 U.S. 1, 7 2011). Justice Ginsburg dissented on the basis that “what is now known about shaken baby syndrome casts grave doubt” on the conviction (2011: 11). The majority dismissed this latter argument as “pure speculation” without examining any of the subsequent research discrediting SBS.

It is perhaps not surprising that a Court concerned with proven innocence would not have much time for residual or lingering doubts. In 2006, the Court ruled that the constitution does not require that an accused be able to argue residual doubt in the capital sentencing stage of a trial. Justice Breyer for the Court stressed that an accused should not be able to introduce evidence that is “inconsistent” with his “prior conviction” Oregon v. Guzek, 546 U.S. 517, 523 (2006). Justice Scalia with Justice Thomas agreed on the merits and argued that residual doubts were in each and every case inadmissible at sentencing. In some ways, this was harsher than Chinese and Indian courts, which, as discussed in Chapters 9 and 10, use residual doubt in some cases to mitigate the death penalty.

The Court has also not revisited any of its guilty plea decisions from the early 1970s that encouraged plea bargaining for efficiency reasons and accepted guilty pleas from accused coerced by the threat of a death sentence into pleading guilty while maintaining innocence (North Carolina v. Alford 400 U.S. 25 1970). The Alford plea is still used in most American jurisdictions. It has been used in cases involving wrongful convictions, most famously in the case of the West Memphis Three, where three men entered an Alford plea to get out of a jail after eighteen years. Keith Findley and his colleagues in an important article on “bargaining away innocence” have documented how prosecutors often offer deep “time served” sentencing discounts to avoid re-trials in cases of suspected innocence and that 59 per cent of accused accept such “bargains” (Findley et al., Reference Findley2022). For many, the promise that the American criminal justice system makes about recognizing innocence is a false promise. This raises serious issues about whether the Court’s rhetoric about innocence only attempts to legitimate a system that continues to convict many innocent people.

6.5.5 Qualified Immunities and American Extralegalism

The Supreme Court’s performance has been no better on the accountability of state actors for wrongful convictions. In a 2009 decision by Justice Breyer, the Court applied the absolute immunity of prosecutors for judicial functions to the actions of a prosecutor who had been responsible for the infamous use of unreliable jailhouse informers that led to many wrongful convictions in Los Angeles. A 1989 Grand Jury report had found a deliberate and informed failure by prosecutors to take actions necessary to prevent jailhouse informers from lying and contributing to wrongful convictions. There was no systemic attempt to keep track of informers who lied. None of this was evident in the Supreme Court’s unanimous ruling, which formalistically posited that prosecutors were absolutely immune for any record-keeping system because it was tied to the calling of evidence at trial (Kamp v. Goldstein, 555 U.S. 335, 346 2009).

Two years later, the Court in a 5:4 decision decided that New Orleans prosecutors were not liable for clear disclosure violations in withholding exculpatory material. Justice Thomas for the majority rejected evidence of four other exonerations on the basis of undisclosed evidence. He reached this conclusion on the legalistic and unconvincing basis that the undisclosed evidence in the other cases was not of the same type as the undisclosed evidence in the plaintiff’s case (Connick v. Thompson, 563 U.S. 51 2011). The Court’s decision overturned the $14 million damage award that John Thompson has received. Thompson, a Black man, had spent fourteen years on death row. He wrote:

Extralegalism (Roach, Reference Roach2011: 164) – the use of law to condone injustice and escape accountability – permeates this and many other decisions. As Jack Goldsmith, who as a Bush administration official ordered that the infamous torture memos be withdrawn, has argued that the memos were “excessively legalistic” and effectively prevented accountability measures, at least in the United States (Goldsmith, Reference Goldsmith2007: 102; Roach, Reference Roach, Huff and Killias2013a). The concept of extralegalism helps explain why prosecutors in the LA jailhouse informer and the New Orleans prosecutor office escaped civil accountability for egregious acts that caused wrongful convictions.

One exception to extralegalism is the over $3.5 billion of awards and settlements that just over 850 exonerees since 1989 have received (Gutman, Reference Gutman2025). But these lawsuits have only benefited about 25 per cent of all exonerees. They require proof of police fault, a concept also used by the Supreme Court to avoid expanding its doctrines protecting the accused from unreliable eyewitness identification (Perry v. New Hampshire, 565 U.S. 228), unreliable confessions caused by mental illness (Colorado v. Connelly, 479 U.S. 157) or unreliable evidence such as that based on SBS (Cavazos v. Smith 565 U.S. 1). American extralegalism is a useful comparative concept because it helps explain why the United States, which has so many lawyers, so frequently engages in conduct that so many in other democracies believe are extralegal (Kagan, Reference Kagan2019; Roach, Reference Roach2011: 436–438).

6.6.1 States Moratoriums and Abolition

The most obvious connection between exonerations and abolition came in the wake of thirteen death row exonerations and a 2002 commission that recommended abolition but also made eighty-five other recommendations designed to prevent wrongful convictions in Illinois (Illinois, 2002). This resulted in a Republican Governor George Ryan imposing a moratorium on all executions in 2000 and a blanket pardon of all 167 people on death row when he was leaving office in 2003. In 2011, Illinois abolished the death penalty, making it the 16th state to abolish the death penalty.

In Maryland, a commission that included Kirk Bloodsworth, the first man exonerated from death row by DNA, recommended the abolition of the death penalty. It cited as reasons the risk of convicting the innocent including limits and the possibility of error in DNA testing combined with issues of cost, delay and discriminatory application of the death penalty. It stressed that “the majority of criminal cases do not include biological evidence that definitively determines the identity of the perpetrator through DNA testing. It is estimated that credible DNA evidence is available in only ten to fifteen percent (10–15%) of death penalty cases” (Maryland, 2008: 20). The next year Maryland rejected abolition, but limited the use of the death penalty to cases where there was DNA evidence, a video linking the accused with the murder or a video-taped confession. In 2013, Maryland abolished its “fool-proof” death penalty. Mitt Romney as Governor of Massachusetts tried but failed to persuade the legislature to re-instate a “fool-proof” death penalty that would only execute people on the basis of scientific evidence (Massachusetts, 2004).

Virginia abolished the death penalty in 2021. Just under half of the American states have abolished the death penalty. Populist concerns about the execution of the innocent have eroded public support for the death penalty in a way that the focus on racial discrimination in the application of the death penalty did not (Baumgartner et al., Reference Baumgartner2008). At the same time, there is no cause for complacency. As will be discussed later, Trump’s rush to execute thirteen people at the end of his first Presidency indicated that there is still support for the death penalty. As will be seen, the death penalty is likely to remain constitutional in the United States. Even when repealed, it is often replaced with harsh sentences of life imprisonment with no possibility of parole.

6.6.2 Unsuccessful Constitutional Challenges

It is striking how elected executives and legislatures motivated by death row exonerations have in the DNA era taken steps towards abolition that have been resisted by a more conservative American judiciary. This suggests that professional judicial concerns about the finality of convictions and deference to state prosecutions have had more sway than revulsion at the prospect of executing the innocent.

A direct attempt to challenge the death penalty based on the risk of convicting the innocent was accepted by one American judge, Jed Rakoff, but quickly overturned on appeal. Judge Rakoff is a scholarly judge best known for his frequent contributions to the New York Review of Books. In his extrajudicial writing, he explained how after his brother was murdered in 1985, he wanted the perpetrator to be executed but came later to change his mind. One reason, not mentioned in his decision, was his recognition that the American death penalty is applied “in a racially unjust manner” with most executions taking place in the South and related to the legacy of slavery and lynching (Rakoff, Reference Rakoff2021: 49). He also concluded that the danger of executing the innocent “is the factor most likely to deprive the death penalty of its moral force and therefore ultimate legal justification” (Reference Rakoff2021: 56).

In 2002, Judge Rakoff raised the constitutionality of the death penalty on his own motion and issued both a preliminary and final decision holding that the death penalty was unconstitutional. He cited mounting evidence that innocent people were more frequently convicted than previously thought. He concluded that the death penalty violated both procedural due process by depriving people of the ability to prove their innocence and substantive due process because of the risk of executing the innocent (United States v. Quinones, 196 F. Supp. 2d 416, 420 S.D.N.Y. 2002; United States v. Quinones 205 F Supp 2d 256. S.D.N.Y. 2002).

Judge Rakoff’s decision was quickly overturned on appeal. The 2nd Circuit in a decision written by Judge Jose Cabranes, a former general counsel for Yale University and appointed to the bench by President Jimmy Carter, ruled that the due process clause itself contemplated the death penalty in its reference to the death penalty (United States v. Quinones, 313 F.3d 49 2002). This textual defence of the death penalty subsequently found favour with Justice Scalia in Glossip v. Gross, 576 U.S. 863 (2015). Four liberal judges in that case would have ruled the death penalty to be unconstitutional in part because of mounting evidence of death row exonerations and a few wrongful executions. Justice Breyer pointed to Texas’s execution of Carlos Deluna (Liebman, Reference Liebman2014) and Cameron Todd Willingham as examples of wrongful executions as well as historical cases of executive pardon (Glossip v. Gross, 576 U.S. 863, 895 (2015)).

None of the American decisions cited a 2001 decision by the Canadian Supreme Court. The unanimous Court ruled that a due process clause in Canada’s 1982 constitutional bill of rights should be interpreted as prohibiting the extradition of fugitives to face the death penalty (US v. Burns, 2001 SCC 3). It overruled 1990 precedents that allowed extradition to face the death penalty. The Court justified this change on the basis of growing death row exonerations in the United States, the decisions by English courts on initial references by the English CCRC that found wrongful executions in the 1950s and Canada’s own experience with wrongful convictions including DNA exonerations. The focus on innocence in the Canadian case has been criticized for not focusing on broader issues of dignity that would apply even to the guilty and that influenced an earlier decision of the South African Constitutional Court to invalidate the death penalty (Jouet, Reference Jouet2022; Martin, Reference Martin2002). The 1995 South African decision did not, however, ignore the risk of wrongful convictions. It quoted with approval a state court decision striking down the death penalty under Massachusetts constitution on the basis that a court “cannot raise the dead” (S v. Makwanyane and Another (CCT3/94) [1995] ZACC 3 at para 54 citing Suffolk District v. Watson, 381 Mass. 648, 663 1980). A concern about the risk of executing the innocent was the most likely basis for constitutional invalidation of the death penalty in the United States. It might have succeeded had more liberal judges been appointed to the United States Supreme Court.

American courts will not constitutionally invalidate the death penalty in the foreseeable future. Trump has and will continue to appoint judges who will follow Justice Scalia’s lead. Justice Scalia argued that all of the academic studies of wrongful convictions were written for abolitionist aims and inflated the number of exonerations. He cited a brutal rape/murder of a child as justification for the death penalty (Callins v. Collins 510 U.S. 1143, 1143, 1994) but one of the accused, Henry McCollom, was subsequently released after 30 years on death row because of DNA. He accepted prosecutor’s Joshua Marquis’s, 2005 conclusion that there was an acceptable 0.027 per cent error rate in the American criminal justice system. This figure was based on the denominator of all 15 million felony criminal convictions between 1989 and 2003 (Kansas v. Marsh 548 U.S. 163 (2006)). The United States’ highly punitive system itself was used to produce a low error rate that was deemed by Justice Scalia to be “an insignificant minimum.” (Kansas v. Marsh 548 U.S. 163, 199 (2006)).

6.6.3 The Trump Executions

After a twenty-year pause on federal executions and after his loss in the November 2020 election, Trump had thirteen people executed: six Black men, five white men, one white woman and one Indigenous man. One Black man, Brendan Bernard, was executed even though at eighteen years of age he only acted as an accomplice by setting fire to a car after he thought the victims were already dead. In several of the cases, executions occurred even though the condemned were still litigating claims in Federal court. The new Biden Administration imposed a moratorium on federal executions soon after coming into office on the basis of “serious concerns” about “its impact on people of colour and the troubling number of exonerations” (National Public Radio, 2021).

Justice Sotomayor filed a dissent from an unsigned opinion by the six-judge Conservative majority authorizing Trump’s 13th execution. She observed: “the Federal Government will have executed more than three times as many people in the last six months than it had in the previous six decades.” The Court often without giving any reasons and making decisions within hours “has consistently rejected inmates’ credible claims for relief … Thereby ensuring those prisoners’ challenges would never receive a meaningful airing.” She also noted that in two of the cases, new evidence about intellectual impairment had been ignored and that a third person had been executed despite evidence that he suffered from Alzheimer’s disease (United States v. Dustin John Higgs 592 U.S. 2021).

Biden, at the end of 2024, pre-empted Trump’s anticipated resumption of federal executions at the start of his second term by commuting the death sentences of thirty-seven of forty people on federal death row, including fifteen Black men and six Latino men. The only exceptions were for those convicted of terror- or hate-motivated murders. In a sign of the growing partisan divide on the issue, Biden explained: “In good conscience, I cannot stand back and let a new administration resume executions that I halted” (New York Times, 2024c). Trump responded with an executive order on January 20, 2025, affirming the need for and the popularity of the death penalty; opposing any judicial impediments to its use and encouraging the states to seek the death penalty for the thirty-seven people pardoned by Biden, who he described as “vile and sadistic.” The executive order also ordered the federal government to assist the states in obtaining chemicals for lethal injections to carry out their death penalty (Executive Order 14164). Executions may increase during Trump’s second term.

6.8.1 Compensation Laws

Despite Borchard’s advocacy for compensation laws for those who could prove that they were innocent and not at fault for their wrongful conviction, only four states and the federal government had enacted such laws by 1976 (Rosenn, Reference Rosenn1976: 725). Since that time, most states have enacted such compensation laws, though a number of states including Arizona, Kentucky and Pennsylvania have not. Pennsylvania has the eighth highest number of exonerations (Bazelon, Reference Bazelon2018a: 42). Some states, including New Jersey and Ohio, bar compensation for those who pled guilty. Georgia enacted a compensation law in 2025, but it only won bipartisan support after it was paired with legislation compensating Trump for his legal expenses for being prosecuted for election racketeering in relation to the 2020 Presidential election.

Compensation statutes are often defended on the basis that they provide more regular compensation with less access to justice barriers than relying upon litigation (Bernhard, Reference Bernhard2004). Some recent findings by Jeffrey Gutman and Lingxiao Sun (Reference Gutman and Sun2018) should temper some of these arguments. Of the first 2,000 wrongfully convicted and listed in the National Registry, they found that under 53 per cent of the exonerees in states with compensation statutes filed claims. There were low filing rates in Florida, which precluded compensation if applicants have a felony conviction. As of 2022, only five of seventy-two exonerations in the registry received statutory compensation in Florida, and only one exoneree qualified after Florida in 2017 slightly loosened its “clean hands” requirement by only excluding from compensation anyone who was convicted of a violent felony or more than one non-violent felony (Gutman, Reference Gutman2022). Claim rates were also low in Missouri, which permits compensation with only exonerating DNA evidence, and in Maryland and Tennessee, which required the Governor to grant a pardon as a precondition to receiving compensation (Gutman and Sun, Reference Gutman and Sun2018: 720, 744). The lowest average awards were an average of $2,623 for each year sixteen exonerees in Wisconsin were imprisoned and $26,785 for each year seventeen exonerees were imprisoned in Mississippi (Gutman, Reference Gutman2023: 3). Federalism allowed for these outlier states.

Gutman and Sun found a 73.5 per cent success rate for statutory claims, 17.5 per cent of claims being denied and the rest still outstanding. The unsuccessful applicants for statutory compensation came mainly from New York and California and were people who were unable to establish their innocence (Reference Gutman and Sun2018: 720). Professor Gutman’s updated statistics show that 60 per cent of exonerees in the thirty-eight states with compensation statutes have filed claims as of the end of 2024, with 76 per cent being successful in their claims. Over $1 billion has been paid out at an average of just over $68,000 per year wrongfully spent in prison (Gutman, Reference Gutman2025). This is a much more generous compensation system than available in England, especially since its 2014 reforms (Law Commission, 2025; Quirk, Reference Quirk, Jasinski and Kremens2023). Most American compensation decisions are made by agencies or courts, whereas, in England, the decision is made by the elected executive.

6.8.2 Civil Litigation

Gutman and Sun (Reference Gutman and Sun2018) found that about 45 per cent of the first 2,000 exonerees listed on the American registry commenced federal civil rights or state tort litigation. The odds of such litigation were three times greater if official misconduct was one of the causes of the wrongful conviction. This makes sense given that civil rights and tort litigation is generally directed at such misconduct. Of the 808 exonerees who engaged in litigation, 55 per cent succeeded on the merits or obtained a settlement. This was a slightly lower percentage than those who sought statutory compensation (Reference Gutman and Sun2018: 699, 772). The average judgment was $3.8 million (Reference Gutman and Sun2018: 772). Most statutory regimes cap damages at $50,000 a year, with Texas having the highest cap at $80,000 and some laws raising the cap to $100,000 for each year spent on death row. In short, litigation is somewhat more risky but when successful results in a much more generous payout.

Between 1989 and the end of 2024, Professor Gutman has estimated that defendants or their insurers have paid out $3.5 billion from successful lawsuits brought by 851 exonerees. The success rate of such litigation has fallen to 50 per cent, but the average civil award has risen to $4.1 million. Successful litigants receive, on average, almost $327,000 for each year wrongfully spent in prison (Gutman, Reference Gutman2025). The wrongfully convicted who can successfully litigate are much more generously compensated than those who claim under compensation statutes.

Another recent study of jury awards in cases brought by those who were wrongly convicted found an average award of $10.3 million in cases decided between 1989 and 2019. The highest award was $56.8 million, and the lowest was $1 (Cohen, Reference Cohen2021). The mean civil litigation case took fifty-one months to resolve but also rendered an average award of $3,600 a day in prison for cases that were litigated and an average of $1,149 a day for cases that were settled (Reference Cohen2021: 705). The estimated litigation costs of these cases were between $1.4 and $2.6 million (Reference Cohen2021: 720), suggesting that, as with other civil litigation, lawyers were profiting.

Civil litigation is more accessible in the United States than in any other place in the world. Why? An unsuccessful plaintiff does not have to pay the defendant’s litigation costs. Lawyers are often willing to take cases with potential million-dollar judgments on a contingency basis, which means that the plaintiff does not have to pay the lawyer if they lose but have to give their lawyer a significant percentage of any award or settlement. In addition, some civil rights laws encourage litigation by allowing successful plaintiffs (but not successful defendants) to receive their attorney’s fee from the defeated party. The 2nd Circuit in one recent case upheld a jury’s $18 million award to two men exonerated by DNA in a rape/murder. The court also awarded almost $5 million to the men’s lawyers for the eight years of litigation (Restivo v. Hessemann 846 F.3d 547 2017). The exceptional American system of adversarial litigation (Kagan, Reference Kagan2019) results in million-dollar payouts for a minority of exonerees. They are the “lucky” “winners,” but they only constitute a quarter of all exonerees (Gutman, Reference Gutman2025).

The effects of civil litigation on exonerees are an under-researched subject. On the one hand, they could be harmed by adversarial litigation that may attempt to blame them for their own wrongful conviction and diminish the value of what they lost. There is no guarantee that any of the lawyers will be trauma-informed. On the other hand, civil litigation can hold some of those responsible, especially the police, who unlike prosecutors do not enjoy absolute immunity, for wrongful convictions. This may also help explain why so many clients of Innocence Projects receive plea offers that allow them to enter an Alford plea and be released from prison, given that such deals will make subsequent civil litigation more difficult (Findley et al., Reference Findley2022). It is undeniable that civil litigation when successful results in much more generous compensation than seeking capped statutory claims. The American wrongful conviction compensation industry is not the social welfare method of loss distribution that Borchard originally wanted (Borchard, Reference Borchard1914). Instead, like much else in contemporary America (Kagan, Reference Kagan2019), it is driven by high-stakes and adversarial litigation that has resulted in over about 850 exonerees receiving $3.5 billion and over 1,300 exonerees receiving a less generous but still significant $1 billion under compensation statutes that exists in most but not all states (Gutman, Reference Gutman2025).

6.8.3 More Holistic Reparation

Some state compensation laws provide for non-monetary forms of compensation. In a nod to rehabilitation after prison, fifteen states provide for free tuition. Less states provide for medical, housing, and employment assistance, which may be more useful. Voluntary groups such as After Innocence also provide supports including free dental care, mental health care and financial planning. Through litigation, state laws and charity, the American approach to compensation for victims of wrongful convictions is the most generous in the world.

One especially creative and holistic approach to compensation was created by the Chicago City Council to respond to the widespread torture of Black men on Chicago’s south side by white police officers under the command of Jon Burge. It came partly as a result of a decision of the United Nations Committee on Torture criticizing unremedied violence by Chicago police against racialized people. It involved $5.5 million in compensation but also an official apology by the mayor, waived tuition for victims and their family members, counselling for torture victims and their families, mandatory education about the torture in the public schools and the creation of a memorial. The approach has been related to both restorative justice and dignity restoration (Baer, Reference Baer2018). This was much less generous than a reported $108 million in settlements and civil verdicts that Chicago has paid to Burge’s victims. At the same time, the restorative approach also avoids the $57 million paid in legal fees in those cases (Taylor, Reference Taylor2022).

Lara Bazelon (Reference Bazelon2018a: 94) has written of the challenges faced by exonerees who are often traumatized and have had their parents die and their children alienated while they were in prison. Exonerees often have no support system, not even a parole officer, when they are released from prison. She argues for a restorative justice approach that brings together all of those affected by a wrongful conviction. Jennifer Thompson, who wrongfully identified Ronald Cotton as the person who sexually assaulted her but subsequently wrote a book with him about their experiences, has founded a charity, Healing Justice, that provides restorative justice for both exonerees and crime victims in wrongful conviction cases. Such projects recognize some of the similarities between survivors of crime and wrongful convictions.

Municipalities have to pay the vast amount of over $3.5 billion that has been awarded in civil rights litigation because other levels of American government enjoy absolute immunities. Chicago is still paying the bills for its many wrongful convictions. It has paid an estimated $750 million since 2000 (Cherone, Reference Cherone2023). Both the time lag between police conduct and exonerations and the self-insuring approach taken by many cities may prevent these high damage awards and settlements from influencing police conduct. A more efficient form of cost internationalization under China’s state compensation law will be examined in Chapter 9.8.

6.9.1 Take on Systemic Misconduct and Mass Exonerations

Innocence projects have not played a key role in the mass exonerations examined in this chapter, with other groups like the NAACP, the ACLU and public defenders taking the lead role (Keever, Reference Keever2023). Innocence projects have expertise and experience with police, prosecutorial and forensic misconduct and false guilty pleas, which would be valuable in redressing many of the types of scandals that have resulted in mass exonerations even if not all of the victims could prove their factual innocence. Focusing on such cases would also move away from the narratives of individual innocence and victimhood and towards a more systemic critique of the criminal justice most notably with respect to systemic racial discrimination. For example, the follow-up to the Rampart scandal in Los Angeles moved from the injustices of the cases to the need for fundamental policing reform (Covey, Reference Covey2012). Such a move has dangers in an era of political polarization, but it is also supported by arguments that those who support wrongful conviction reforms already tend to be more liberal (Norris, Hicks and Mullenix, Reference Norris, Hicks and Mullenix2023). Focusing more on mass exonerations, especially in drug cases, would also lead to an increased focus on anti-Black racism, given that Black people are nineteen times more likely to be wrongfully convicted of such crimes than white people (Gross et al., Reference Gross2022). This could possibly also lead the innocence movement to support some forms of drug decriminalization, albeit again perhaps at the cost of bipartisan support.

6.9.2 Take on Imagined Crimes That Did Not Happen

The American innocence movement has often focused on the wrong perpetrator paradigm of wrongful convictions. Writing in 1959, Stanley Earle Gardner, founder of the Court of Last Resort, concluded that “it is a great a miscarriage of justice when a guilty person is acquitted as when an innocent man is convicted” (Gardner, Reference Gardner1959: 22). A former Attorney General of Ohio and death penalty supporter who admits he would be unable to defend someone “I knew to be guilty” (Petro and Petro, Reference Petro and Petro2010: 103, 165) has argued that DNA testing does not dishonour a victim and is required by “public safety” to stop real perpetrators from escaping justice (Petro and Petro, Reference Petro and Petro2010: x). His book features chapters on “the devil cheats justice” and “proving guilt” of the real perpetrator after a DNA exoneration. Although such crime control arguments may be appropriate for prosecutors who in the United States generally have to be elected and are supported by public opinion polls that suggest that many people are as concerned about wrongful acquittals as wrongful acquittals (Garrett and Mitchell, Reference Garrett and Mitchell2023), they are increasingly problematic.

Innocence organizations should avoid using inaccurate but politically popular claims that every wrongful conviction is a crime control loss because it allows the guilty to go free. This may be true in many DNA exonerations involving murders and sexual assaults, but the American registry, as of the end of 2024, reveals that 40 per cent of exonerations involve cases where no crimes were committed. Innocence organisations need to educate the public about these “no crime” cases (Henry, Reference Henry2020; Roach, Reference Roach2023a). This may also lead to increased attention about how racialized and marginalized people including women are especially vulnerable to suspicions and stereotypes that can make a crime out of an accident. Such a focus could also help confront critics who seem to assume that wrongful conviction reforms will inevitably produce wrongful acquittals (Cassell, Reference Cassell2018; Marquis, Reference Marquis2005).

6.9.3 Support Abolition of the Death Penalty

Innocence projects could also make clear their opposition to the death penalty in all cases. There is already some positive movement in that direction with Innocence Projects co-founders Peter Neufeld and Barry Scheck and Centurion Ministries founder Jim McCloskey successfully urging President Biden at the end of 2024 to commute the sentences of those on federal death row. Their letter understandably focused on death row exonerations, but also stated that every person “caught up in a capital punishment system, whether state or federal, fraught with endless procedural bars, mind-numbing technical litigation, and emotional rollercoaster rides. It wreaks pain, fear, and anger on the families of victims, defendants, witnesses, and everyone connected to the process. For what purpose?” (Scheck, Neufeld and McCloskey, Reference Scheck, Neufeld and McCloskey2024).

There are other positive signs of the expansion of the American innocence movement. The Innocence Projects have joined with others including defence lawyers to try to expose and reduce the trial penalty, which encourages false guilty pleas and leads to very high guilty plea rates (Brown and Neufeld, Reference Brown and Neufeld2021: 279). Rebeca Brown and Peter Neufeld have also indicated that the Innocence Project hopes “to target the vast misdemeanour system, which overcriminalizes conduct as a means of extending power over marginalized populations” (Reference Brown and Neufeld2021: 280). They have also expressed concerns about the unregulated use of forensic genetic genealogy and facial recognition (Reference Brown and Neufeld2021: 287). They have also expressed concerns about false guilty pleas including low standards of mental competence for a valid guilty plea (Dusky v. United States 362 U.S. 402, 1960). Innocence organizations also have skills that could address suspicious deaths in police custody and the need for bail reform if only to reduce false but perhaps rational guilty pleas.

6.9.4 Take on Broader Miscarriages of Justices

Former Innocence Project lawyer and wrongful conviction scholar Valena Beety (Reference Beety2022: 189–192) has argued for a focus on manifest injustice rather than proven innocence. She has suggested the ancient writ of coram nobis used to overturn convictions of Japanese Americans interned during World War II could serve as an important remedy in achieving this end. Professor Beety has also argued for less of an emphasis on the immediate causes of wrongful convictions and more emphasis on “the hallmarks of mass incarceration: racism, discrimination based on prior involvement with police and the court and system, and no avenues for relief after conviction. These are the hallmarks of manifest injustice” (Reference Beety2022: 209). She has also written: “lawyers alone cannot be a movement…It takes a community” (Reference Beety2022: 275). She has also worked with colleagues to create a guide for litigating miscarriages of justice beyond factual innocence that includes wider justice claims including cumulative error claims as allowed in 27 states, mass exonerations and patterns of police misconduct and racial bias claims as responses to increasing resistance to innocence claims (Beety, Newirth and Thompson, Reference Beety2022). More fundamentally, Beety and her colleagues argue that all accused are entitled to constitutional safeguards “not just the ones where a narrative can be created around a perfect defendant” (Reference Beety2022: 9).

The innocence movement has considerable experience and expertise in exposing and seeking accountability for police misconduct. This could be applied to attempts to challenge police violence. Valena Beety (Reference Beety2021) has argued that victims of police violence are often excluded from the victim compensation scheme because many such statutes, like wrongful conviction compensation statutes, require proof of innocence. Such an approach fits into international approaches to victims’ rights (Roach Reference Roach1999a), which recognize that victims of state power are equally deserving as those of private crime. At the same time, such a broadening of approach would run counter to the advice of some who counsel the innocence movement to avoid appearing anti-police (Gould, Reference Gould and Zalman2014).

6.9.5 Take on the Need for Jury Diversity

Bryan Stevenson’s Equal Justice Initiative (2025) has raised concerns that discriminatory uses of peremptory challenges result in the under-representation of Black people on American juries. It has documented how prosecutors have been trained to use supposedly racially neutral reasons for excluding Black people from juries. It also reviewed studies suggesting that more diverse juries would be less likely to convict Black people or sentence them to death. The report details cases of people eventually exonerated of capital crimes who were convicted and sentenced to death by all-white juries or juries with only one or two Black people even in counties with significant Black populations. For example, Glen Ford was wrongfully convicted of murder by an all-white jury in Louisiana in a county with a population that was 38 per cent Black. The prosecutor subsequently admitted to using peremptory challenges because of his belief that Black people would not agree to a death sentence. Ford served thirty years on death row and died at sixty-five years of age, a year after his exoneration. The Equal Justice Initiative (2025) also recounted the case of Curtis Flowers who was convicted three times by non-diverse juries and served twenty-four years before a Supreme Court ruling responded to the prosecutor’s overt and unapologetic discriminatory use of peremptory challenges to exclude Black jurors in a Mississippi county with a 45 per cent Black population (Flowers v. Mississippi 588 U.S. 284 2018). Curtis Flowers was eventually released on innocence grounds. Innocence organisations could make the abolition of peremptory challenges and other attempts to improve the diversity of juries a priority.

6.9.6 The Challenges and Opportunities of Increased Political Polarization

Trump’s 2024 re-election and increased political polarization provide clear dangers for the innocence movement. But it also provides opportunities.

Innocence denials and the use of the death penalty will likely increase under Trump’s influence. In 2024, Missouri executed a Black Muslim man Marcellus Khalifah Williams for murdering a white woman. A local Democratic prosecutor joined with Williams’ lawyers in a motion to vacate Williams conviction on actual innocence grounds. They stressed there was no forensic evidence linking Williams to the murder with footprints, fingerprints and hair found at the crime scene excluding Williams. They warned that the two witnesses who testified against Williams gave inconsistent evidence and were motivated by reward money. They sought to introduce new evidence not heard at trial that one of the victim’s possessions found in Williams car had actually been given to him by one of the incentivized witnesses. They also relied on DNA found on the murder weapon that was not William’s. Alas, DNA testing, however, revealed that the donor of the DNA was the prosecutor and the investigator who, unlike Williams defence lawyer, did not always wear gloves when touching the weapon. The Missouri courts subsequently ruled that Williams had not proven that the police and prosecutor had acted in bad faith in contaminating the evidence. They also held that the prosecutor’s explanation for using a peremptory challenge to remove a potential Black juror on the basis that the prospective juror looked just like Williams was not intentional racial discrimination. Finally, the Missouri Supreme Court concluded that “there is no clear and convincing evidence that Williams is actually innocent.” (Prosecuting Attorney v. Williams 2024 Mo. Lexis 306 Supreme Court of Missouri). The Missouri Supreme Court split on ideological lines. The United States Supreme Court denied relief on the day that Williams was executed by lethal injections over the dissents of the three Justices on the Court appointed by Democrats (Williams v. Missouri 2024 U.S. LEXIS 3052).

Missouri Governor Mike Parson refused to pardon Williams, who had become an Iman and a poet during his more than twenty years in prison even though, as discussed earlier, he had pardoned a white couple who had drawn guns on Black Lives Matter demonstrators. Trump praised Parson for having “the courage” to pardon the white St. Louis lawyers because in Trump’s words they were “defending their property and if they had not done what they did, their property would have been completely destroyed and they would have been badly beaten, or dead – great going Mike!” (Missouri Times, 2025). Parson also commuted the sentence of a white police officer who was convicted of manslaughter for fatally shooting a Black man (New York Times, 2024e). Despite declines in support and the use of the death penalty during the innocence movement, the death penalty is alive and well in some parts of the United States. Mr. Williams was the 100th person executed by the state of Missouri since 1989 and the fourth in 2024 (Shelly, Reference Shelly2025).

Williams’ execution demonstrates American extralegalism. There was extensive litigation in the case that nevertheless resulted in the state killing a man despite many doubts about his guilt. The execution also reflects political polarization. A former Missouri governor had appointed three judges to investigate whether Williams was innocent, but this investigation was terminated by Governor Parson after he was elected with support from Trump. State Attorney General Andrew Bailey, also supported by Trump, opposed the local prosecutor’s claims that Williams was innocent. He also opposed a last-ditch attempt to save Williams’s life through an Alford plea. Bailey has intervened in other cases where local prosecutors supported innocence claims. He also attempted to stop the court-ordered release of a prisoner on innocence grounds who had been imprisoned for forty-three years (Sarat, Reference Sarat2024).

Innocence is a severe rationing of justice, but it may not always be enough in the United States in the Trump era. This creates conditions for the American innocence movement to focus not only on innocence and the immediate causes of wrongful convictions, but deeper structural factors, including anti-Black racism, police misconduct and the death penalty. The Trump era of white nationalism and political polarization should make it clear that previous struggles for bipartisan consensus that have restrained the innocence movement are no longer sustainable. They should recognize that a significant constituency in the United States, like Trump, will never fully accept the innocence of those such as the Exonerated (Central Park) Five even in the face of DNA exonerations. Trump’s vile actions from his 1988 ads calling for the then teenagers to be executed to his comments in the 2024 Presidential election debate will be examined in Chapter 7.5.

A more aggressive approach that eschews attempts at bipartisan consensus will cost the American innocence movement some political and financial support. At the same time, more recent public opinion data suggests that more liberal Americans are the ones who know and are most concerned about wrongful convictions (Norris, Hicks and Mullenix, Reference Norris, Hicks and Mullenix2023). The illiberalism and racism of the Trump era could help the American innocence movement reveal and defend its liberal and anti-racist soul.