2.2.1 Clive Walker’s Rights-Based Definition

In an influential but controversial definition, University of Leeds Law Professor Clive Walker (Reference Walker, Walker and Starmer1999: 33) defined miscarriages of justice as:

This is a very broad definition of miscarriage of justice that essentially extends to all state action that violates human rights. It was shaped by Professor Walker’s expertise in counterterrorism and his extensive knowledge not only of wrongful convictions in Northern Ireland but also of the use of internment, preventive detention, restrictions on freedom of expression and even official complicity in violence by paramilitaries (Roach, Reference Roach2019c). In a review of Walker’s broad definition, Richard Nobles and David Schiff (Reference Nobles and Schiff1994: 384) argued that Walker (Reference Walker, Walker and Starmer1993), “by defining miscarriages of justice to include unjust laws and failure to protect individuals against actions taken by criminal and vigilantes,” had gone well beyond the purview of wrongful convictions.

In earlier work, I shared some of Nobles and Schiff’s misgivings about the breadth of Walker’s definition, while at the same time arguing that miscarriage of justice should be extended to include the unwarranted use of military and immigration detention in the United States after the 9/11 terrorist attacks (Roach and Trotter, Reference Roach and Trotter2005). Subsequently, I have come to appreciate Walker’s rights-based definition both as an academic definition (Roach, Reference Roach2019c) and something that could be harnessed to the international law right to an effective remedy for all human rights and one that could, following a legal process approach, be delivered by courts, legislature or the executive (Roach, Reference Roach2021).

My only remaining reservation about Walker’s definition is its application to “whenever the rights of others are not effectively or proportionately protected or vindicated by State action against wrongdoers” (Walker, Reference Walker, Walker and Starmer1999: 33). Brian Forst (Reference Forst2004: ch 3) calls these “errors of impunity.” Like Walker, Forst (Reference Forst2004: 17–18) accepts a broad definition of a miscarriage of justice that includes “the failure to hold culpable offenders for the crimes they commit.” I am not troubled at calling lynching and other forms of murder openly condoned by the state a miscarriage of justice. Moreover, I have even argued that “wrongful acquittals” that are tainted by state-sanctioned discrimination can also be seen as a miscarriage of justice (Roach, Reference Roach2019b: ch 7).

That said, I am uncomfortable with Professor Forst’s (Reference Forst2004) suggestions that all justice errors – whether of wrongful convictions or wrongful acquittals – are equal and should be balanced as such. The English Law Commission (2025: 4.14) has also recently and unhelpfully suggested that a miscarriage of justice is the acquittal of the factually guilty at least where it results from legal error and misconduct, as opposed to the reasonable doubt principle. Paul Cassell’s (Reference Cassell2018: 853–855) argument that one is much more likely to be a crime victim than a truly innocent victim of a wrongful conviction is correct, but that does not suggest that societies are overinvesting in preventing wrongful convictions (Cassell, Reference Cassell1998). Ronald Allen and Larry Laudan (Reference Allen and Laudan2008: 83) dramatically but unhelpfully pose the question as whether it is better to be wrongfully convicted of murder or murdered. A straight-up balance between wrongful convictions and wrongful acquittals oversells the ability of criminal prosecutions to prevent crime. It posits fully effective crime control even though this is much less likely to be achieved than the avoidance of state-inflicted miscarriages of justice (Roach, Reference Roach1999b). The arguments made by Professors Forst and Cassell also elide the state’s more direct role and responsibility when detaining and imprisoning people as opposed to ensuring that they do not suffer crimes, even if those crimes are conceived as threats to human rights.

Walker’s definition may avoid some of these concerns in its reference to the need for the state to act proportionately to protect the rights of crime victims. As suggested in Chapter 1 and in some other work (Roach, Reference Roach2006, Reference Roach2021), the concept of the state having to justify limits it places on the rights as proportionate and necessary serves as a principled and transparent basis for factoring in important and legitimate interests without creating an undisciplined and open-ended form of balancing. Walker’s definition makes clear that the rights of both suspects and potential victims can be subject to proportionate and justified limits. Still, Walker’s incorporation of effective crime prevention in protecting the rights of crime victims through “State action against wrongdoers” (Walker, Reference Walker, Walker and Starmer1999: 33) might still stretch the capacious term “miscarriage of justice” beyond its breaking point. But in all other matters, I find Walker’s rights-based approach to be normatively attractive.

Writing in 2014 with Carole McCartney, Clive Walker made clear that miscarriages of justice were not limited to those who could be proven to be factually innocent but includes all those “unjustly convicted” (McCartney and Walker, Reference McCartney and Walker2014: 22). As I have explained elsewhere, it is mistake to think that Walker’s definition of a miscarriage of justice does not also include the conviction of the innocent (Roach, Reference Roach2019c: 192). Walker specifically includes cases where “there is no factual justification” for either their punishment or treatment. It is important to understand and to be clear that miscarriages of justice is a broad term that includes, but is not limited to, both wrongful convictions and the wrongful convictions of those who subsequently prove their innocence.

2.2.2 Miscarriages of Justice as a Ground for Quashing Criminal Convictions

When the English Court of Criminal Appeal was created in 1907 in response to a few high-profile wrongful convictions, it was given a mandate to overturn a conviction if “the verdict of the jury should be set aside on the ground that it is unreasonable or cannot be supported having regard to the evidence … or that on any ground that there is a miscarriage of justice” (Criminal Appeal Act, 1907 s.4(1)). At the same time, the Court of Appeal could affirm a conviction, despite errors of law, if there was “no substantial miscarriage of justice.” The concept of miscarriage of justice was central to the mandate of the new Court of Criminal Appeal.

The original appeal provisions were transplanted to other countries, including Australia, Canada, Scotland and New Zealand, where appellate courts still allow appeals from convictions on grounds related to miscarriages of justice (Sangha, Roach and Moles, Reference Sangha, Roach and Moles2010). The term miscarriage of justice has defied a comprehensive definition by appellate courts. As Professors Nobles and Schiff have concluded that “it is impossible to capture the legal meanings of miscarriages in any particular form of statutory words” (Nobles and Schiff, Reference Nobles and Schiff2000: 55).

The New Zealand Supreme Court has defined miscarriage of justice as an important catch-all and fail-safe ground of appeal that “can potentially apply to anything … which has gone wrong with the substance or process of the case and has not been cured or become irrelevant to the verdict.” It has elaborated that miscarriages of justice “can include something which has occurred either before or during the trial… It must also be taken to include situations where admissible defence evidence is wrongly excluded or where after the trial fresh and cogent evidence comes to notice and casts doubt on the guilty verdict” (R. v. Matenga [2009] 3 NZLR 145 at para 11). This definition of miscarriage of justice incorporates concerns about reasonable doubt about guilt.

The Supreme Court of Canada has stated that a miscarriage of justice for the purpose of an appeal occurs whenever “the error, whether procedural or substantive, led to a denial of a fair trial” (R. v. Khan [2001] 3 SCR 823 at para 27). In 1937, the Australian High Court indicated that a miscarriage of justice includes both cases “where there is affirmative reason to suppose the appellant is innocent” but also where “it appears unjust or unsafe to allow the verdict to stand…” (Davies and Cody v. The King (1937) 57 CLR 170 at 180). The Scottish courts allow appeals from convictions if there was a miscarriage of justice and have stated it “would be difficult or impossible and unwise to attempt a comprehensive definition of the concept of miscarriage of justice” but that it covers “a wide variety of situations in which, for one reason or another, the court concludes that justice has not been done, in the particular circumstances of a case” (Harper v. HM Advocate 2005 SCCR 245 at para 33). The English courts have similarly indicated that “if for whatever reason, the courts concludes that the appellant was wrongly convicted, or is left in doubt, whether the appellant was rightly convicted of the offence or not, then it must of necessity consider the conviction unsafe” (R. v. Graham [1997] 1 Cr. App. R. 302, 308).

Miscarriages of justice as a traditional ground of appeal include both factual and legal errors. They include, but are not limited to, cases where there are concerns about innocence. A miscarriage of justice and the closely related English ground of safety do not require proven innocence. Even though accused may formally have the burden on appeal to overturn a conviction, many appeal-based definitions of miscarriages of justice incorporate in some ways the fundamental principle in a liberal criminal law that the accused should be given the benefit of a reasonable doubt. That said, the term miscarriage of justice remains vague and slippery. It will be suggested later that the flexibility of the term provides a basis for courts, often covertly, to balance their concerns about injustice with competing concerns about the finality of convictions and the limits of their appellate role (Nobles and Schiff, Reference Nobles and Schiff2000).

2.2.3 Miscarriage of Justice as a Ground for Compensation

In interpreting a provision for compensation, the Irish courts have indicated that a miscarriage of justice includes cases where prosecutions should never have been brought because of the lack of credible evidence implicating the accused or cases that involve a “grave defect in the administration of the justice” including departures from rules that made a trial “irreconcilable with judicial or constitutional procedure” (DPP v. Wall [2005] IECCA 140). A miscarriage of justice also includes cases involving “the concealment, or material non-disclosure, of witness statements focused on a central issue in the prosecution case which tend to support an actual defence for the accused” (Abdi v. Director of Public Prosecutions [2022] IESC 24 at para 46). It also occurs if there is “a substantial and fundamental failure in the trial process and one which undermines in a significant way the prosecution case as accepted by the jury” (Abdi: at para 47).

The Irish Supreme Court helpfully explained that it would maintain its broad approach to compensating miscarriages of justice even if the relevant international law standards for compensation required proof of innocence. It stated:

The Irish approach to compensation for miscarriages of justice does not require proof of innocence. It allowed for compensation of a person who served sixteen years in prison for killing his infant but was subsequently found not guilty by reason of insanity as a result of new psychiatric evidence (Abdi [2020] IEHC 434.)

2.2.4 Miscarriages of Justice and Unwarranted Pre-trial Detention

A number of European jurisdictions provide compensation not only for wrongful convictions but also for wrongful pre-trial detention that results in an acquittal or a decision not to proceed with a prosecution. Italy paid €819 million to 30,017 people who were wrongly detained between 1992 and 2021 compared to €76 million to 214 people who were wrongly convicted during that same time period (Lonati, Reference Lonati, Jasinki and Kremens2023: 52, 54). This underlines how miscarriages of justice related to unlawful detention may be more prevalent than those related to wrongful convictions.

A focus on unlawful imprisonment is appropriate, especially in the Global South, where many people serve long periods in pre-trial detention. One example of a broader and more holistic approach to unwarranted detention is found in Article 46 of Algeria’s 2020 Constitution. It provides that “[a]ny person who is the subject of arbitrary detention, arbitrary temporary confinement, or miscarriage of justice shall have the right to compensation” (Roach, Reference Roach2024a: 744–745). This combines respect for Article 9(5) of the ICCPR, which provides a right to compensation for an unlawful or arbitrary detention, and Article 14(6), which provides a right to compensation for some wrongful convictions. It also follows the Irish Supreme Court’s observation by going beyond the minimal standards as protected by international law.

2.2.5 Miscarriage of Justice and Sentences

The Criminal Cases Review Commissions created in England, Scotland, Norway and New Zealand allow those convicted at trial to apply for relief after appeals are exhausted not only with respect to convictions but also with respect to sentences. They apply tests related to the likelihood that courts would, in light of new evidence not previously considered, change the sentence. This reflects a basic concern about the injustice of unlawful imprisonment, including even unlawful imprisonment for a few days because of technical errors in deducting time spent in pre-trial detention from the accused’s sentence. These four countries recognize that even one day of unlawful imprisonment is an injustice caused by the state. The first case that the New Zealand Criminal Cases Review Commission referred back to the courts involved new evidence that the applicant was fifteen years of age when he was sentenced. The courts accepted this new evidence and held that they had erred by sentencing the applicant as if he were an adult (G v. New Zealand Police [2024] NZHC 189).

North Carolina’s Innocence Inquiry Commission, created in 2006, and Canada’s Miscarriage of Justice Review Commission, created in 2024, do not allow applicants to apply with respect to sentences. Concerns were expressed in Canada that the commission should reserve its scarce resources for considering wrongful convictions. This was done, despite arguments that a “manifestly excessive custodial sentence is also a troubling injustice” (Roberts and Azmeh, Reference Roberts and Azmeh2023: 519) and arguments that new facts were often needed to place the sentencing of Indigenous offenders in proper context (LaForme and Westmoreland Traore, Reference LaForme and Westmoreland-Traore2021; Pate, Reference Pate2022). The omission of sentencing from the jurisdiction of the North American commissions reflects rationing of justice and punitive penal cultures (Garland, Reference Garland2001).

2.3.1 Early Doctrinal Writing about Wrongful Convictions

In 1897, Maurice Lailler and Henri Vonoven wrote an impressive 600-page book called “Les Erreurs Judiciaries et Leurs Causes.” It featured detailed case studies of over ninety cases that constituted “judicial errors.” What is noteworthy, however, is that Lailler and Vonoven did not limit wrongful convictions to those who were clearly innocent. Rather, they defined wrongful convictions as cases corrected by either judicial or political authorities and stressed that it was a judicial error for “le juge criminel, en effet, ne doit pas condamner s’il ne peut appuyer sa conviction sur une certitude absolue. Si la culpabilite n’est pas d’une evidence telle que’innocence apparaisse comme impossible, le magistrat doit se refugier dans le doute et clore le proces par un acquittement” (Lailler and Vonoven, Reference Lailler and Vonoven1897: 3). They stressed that victims of judicial error may be factually innocent or guilty, but they were entitled to be acquitted because “incertitude profite a l’accuse” (Reference Lailler and Vonoven1897: 4). To hammer home this point, they added: “la prosomption d’innocence est la loi dominante de l’instruction criminelle. Les philosophes, les legislateurs, les juristes l’ont formule de tout temps.”

In 1911, the German lawyer and academic Erich Sello wrote a book that focused on factual innocence. Its translated title was The Errors of Criminal Justice and their Causes: A History of Judicial Murder from 1798 to 1910 (Grunewald, Reference Grunewald2023: 31). Sello focused on cases of innocence in order to support his provocative thesis that capital punishment should be abolished to prevent “judicial murders.” At the same time, he recognized that establishing either innocence or guilt to a certainty was often impossible.

Another German scholar and lawyer, Max Hirschberg built on the work of Lailler and Vonoven. Unlike Sello, Hirschberg did not limit his study of wrongful convictions to those who could be proven to be factually innocent. For example, he included as wrongly convicted those who committed a criminal act but should have had a defence. He saw the jury’s conviction and subsequent execution of Charles Guiteau for the assassination of US President Garfield as a wrongful conviction because of Guiteau’s mental disorder (Hirschberg, Reference Hirschberg1940: 40, Reference Hirschberg1941). Like Lailler and Vonoven, Hirschberg stressed the presumption of innocence. He even went beyond the reasonable doubt standard of proof when he argued: “every doubt, and not merely ‘reasonable’ doubt, should be overcome … Not probability but certainty should be demanded” at least in death penalty cases (Reference Hirschberg1941: 46).

2.3.2 Wrongful Convictions and Article 14(6) of the ICCPR

Although it refers to miscarriages of justice, Article 14(6) of the ICCPR provides a right to compensation for quite a narrow subset of all wrongful convictions. It provides:

The reference to “final decision” and a conviction either being reversed or pardoned on the basis of a “new or newly discovered fact” suggests that this provision does not encompass wrongful convictions that are corrected before the final appeal. The right to compensation also does not apply if it is proved that the wrongfully convicted was at fault for the non-disclosure of the conclusive exonerating evidence. In Chapter 11, it will be argued that the restrictive nature of Article 14(6) is related to American and English resistance to it during the drafting of the ICCPR in the 1950s and that the right should be updated given new knowledge about false confessions and false guilty pleas.

Article 14(6) is restrictive, but it stops short of requiring proven innocence. The South African law professor Jamil Mujuzi (Reference Mujuzi2019) and the British judge Lord Phillips have made strong arguments that Article 14(6) does not require proof of innocence as a precondition for compensation. They point to the defeat of a 1959 Israeli proposal near the end of the drafting of the ICCPR that would have provided that “[a] person proved innocent of a crime of which he had previously been convicted, shall have the right to compensation for any damage or punishment which he suffered through no fault of his own” (Clooney and Webb, Reference Clooney and Webb2021: 241). Lord Phillips has concluded that the vote against the Israeli proposal while providing “no positive indication of precisely what the state parties intended ‘miscarriage of justice’ to mean” does make “it difficult to argue that they intended it to mean ‘conviction of the innocent’” (Adams v. Secretary of State [2011] UKSC 18 at [20]).

2.3.3 The Changing Standards of Section 133 of the UK’s Criminal Justice Act, 1988: From Miscarriage of Justice to Proven Innocence

The most developed jurisprudence implementing Article 14(6) of the ICCPR comes from the English courts when interpreting s.133 of the Criminal Justice Act, 1988. The English courts rejected the broader view taken by the Irish courts that a “grave defect in the administration of the justice” and acts that are “irreconcilable with judicial or constitutional procedure” (DPP v. Wall [2005] IECCA 140) are miscarriages of justice that merit compensation. For example, the House of Lords rejected compensation for a person who was subject to extraordinary rendition to face terrorism charges (R v. Sec of State ex parte Mullen 2004 UKHL 18). In 2011, the UK Supreme Court restricted the meaning of a compensable miscarriage of justice even more by concluding that it did not include unsafe convictions (R (Adams) v. Secretary of State for Justice [2011] UKSC 18 at paras 55, 94, 114). Thus, the English courts maintained a safety standard for hearing appeals from criminal convictions while rejecting the same safety standard as a basis for compensation. It rationed funds paid for compensation more restrictively than post-conviction relief on appeal.

Even as the courts restricted their definition of miscarriages of justice to limit compensation, they rejected arguments made by Lord Steyn in dissent that only “the clearly innocent” should be compensated (R v. Sec of State ex parte Mullen 2004 UKHL 18 at para 56). Four judges in dissent in Adams would have required proof of innocence. Lord Brown defended a proven innocence test as necessary to “ensure that a guilty defendant is not compensated for the consequences of his conviction,” something that would appall “the man in the street” (R (Adams) v. Secretary of State for Justice [2011] UKSC 18 at para 277). At the same time, he candidly conceded that the proven innocent test would exclude from compensation those who are innocent but lack the ability to prove it beyond a reasonable doubt (R (Adams) v. Secretary of State for Justice [2011] UKSC 18 at para 281).

In 2014, Parliament amended the compensation scheme to require proof beyond a reasonable doubt that the accused did not commit the offence. Then, Home Secretary Teresa May defended the change as ensuring that “compensation is paid only to those who are clearly innocent” (Quirk and King, Reference Quirk, King, Lennon and McCartney2019: 278). These amendments were criticized by the Joint Parliamentary Committee on Human Rights as a violation of the presumption of innocence (Joint Committee on Human Rights, 2013: ch 5, para 156). Nevertheless, in an England heading towards Brexit and under the leadership of Prime Minister David Cameron, the 2014 amendments went through with only minor and cosmetic changes (Anti-Social Behaviour, Crime and Policing Act, 2014 s.175 amending s.133 of the Criminal Justice Act, 1988). As Hannah Quirk and Colin King (Reference Quirk, King, Lennon and McCartney2019: 278) have argued, this change was “reflective of increasingly punitive and populist attitudes” encouraged by both Labour and Conservative governments in the UK. In short, a country’s interpretation of a miscarriage of justice can reflect its penal culture.

2.3.4 Registries of Wrongful Convictions

As discussed in Chapter 1, registries of wrongful convictions prepared by University researchers in the United States, the UK, Europe, Chile and Canada are an important new source of information. The lists compiled by the innocence organizations, including the Innocence Project co-founded by Barry Scheck and Peter Neufeld, stressed the factual innocence of exonerees. This made sense given that, as Michael Naughton (Reference Naughton2013) has argued, factual innocence is a lay or populist concept, and such projects rely on charitable funding. At the same time, in their academic writings, Scheck and Neufeld were always careful to stress that DNA exonerations should not be the litmus test for wrongful convictions and that they often depended on the luck of what evidence was retained. Scheck and Neufeld predicted that DNA exonerations would eventually end assuming investigative competence in the minority of cases that involve DNA evidence (Scheck, Neufeld and Dwyer, Reference Scheck, Neufeld and Dwyer2000).

The American and self-styled National Registry of Exonerations defines an exoneration as any case where, after a conviction, a person “was relieved of all the consequences of the criminal conviction” either by being declared by a government official to be “factually innocent” or through an acquittal or dismissal of charges after new “evidence of innocence became available.” In addition, “The evidence of innocence need not be an explicit basis for the official act that exonerated the person. A person who otherwise qualifies has not been exonerated if there is unexplained physical evidence of that person’s guilt” (National Registry, Glossary: Exoneration). The Canadian Registry of Wrongful Convictions takes a similar approach but with less references to innocence. It defines a wrongful conviction as a case where the accused is convicted and subsequently is acquitted or has charges not pursued on the basis of new evidence relevant to guilt or innocence. It can also include cases with the indicia of wrongful convictions (Canadian Registry, “What is a wrongful conviction?”; Roach, Reference Roach2023a). Unlike the American and European registries, however, the Canadian registry uses the term wrongful convictions as opposed to the term exoneration because exoneration is an extralegal term (Roach, Reference Roach and Beare2008) often associated with factual or proven innocence. American commentators have focused on the word exoneration as a reason for denying innocence (Marquis, Reference Marquis2005). Canadian courts have decided that they have no jurisdiction to make declarations of innocence as part of a criminal appeal (R. v. Mullins-Johnson 2007 ONCA 750 at [25]). In a passage cited with approval by Lord Phillips in Adams [2011] UKSC 18 [45], the Ontario Court of Appeal concluded that a “third verdict” of factually innocent would devalue the meaning of a not guilty verdict and the presumption of innocence.

The registries’ approach to wrongful convictions is narrower than some definitions of miscarriages of justice (Naughton, Reference Naughton2007) because they generally do not include convictions that are overturned because of a rights violation either in the investigation or the trial. At the same time, they are broader than proven innocence approaches required under some mostly American laws and by most Innocence Projects. The registries are generally focused on measuring legal errors related to guilt and innocence as determined by the legal system. Their approach is “internal” to the legal system (Sumner, Reference Sumner2024: 33). They also take a legal process approach, as they do not require an acquittal by the courts, but also accept a decision of a prosecutor not to proceed with a retrial and the decision of the executive to award a pardon based on concerns about guilt.

The American law professor Richard Leo (Reference Leo2017a: 70–71) has expressed concerns that substituting the Registry’s legal concepts for factual innocence may undermine “the extraordinary power of innocence to effect meaningful criminal justice reform” and “dilute the emotional and moral valence of what animated the Innocence Movement in the first place the pure tragedy of wrong persons errors and their terrible consequences.” At the same time, the registries are all run by university-based researchers who do not have the same dependence as innocence organisations for charitable funding from the public. Moreover, they are well placed to reveal larger patterns in wrongful convictions. Both the American and Canadian registries, for example, have revealed significant percentages of wrongful convictions based on imagined crimes that did not occur at all. This reveals the fallacy of Professor Leo’s equation of genuine innocence cases as “what used to be called wrong man errors” (Reference Leo2017a: 60). Professor Leo is correct, however, to conclude that the American registry has “moved away from actual or factual innocence (what use to be called wrong man errors) to a more legalistic idea of an ‘exoneration’ that, as defined by the Registry, is essentially an erasure of a pre-existing conviction by a governor, prosecutor, judge or jury based on some new evidence of factual innocence” (Reference Leo2017a: 60).

2.4.1 Edwin Borchard: A Pioneer and Entrepreneur of the Proven Innocence Standard

Edwin Borchard was a wide-ranging scholar who is perhaps best known for his 1932 book Convicting the Innocent: Errors of Criminal Justice, which outlined sixty-five cases of “completely innocent people” being convicted (Borchard, Reference Borchard1932: xii). This book is rightly credited for establishing the path of much subsequent wrongful conviction scholarship (Leo, Reference Leo2005). Borchard stressed that he only included cases of “completely innocent people” (Borchard, Reference Borchard1932: xii). From a legal process perspective, the innocence of those he included in the book was so clear-cut that most of them had received remedies from the elected executive or legislature, a fact cited by the majority of the United States Supreme Court as evidence that executive clemency would prevent the execution of the innocent (Herrera v. Collins 506 U.S. 390 at 415 (1993)). As Marvin Zalman (Reference Zalman2020) has noted, Borchard went out of his way to avoid controversy in compiling his list of the completely innocent. For example, he did not include the Italian anarchists Sacco and Vanzetti even though in private correspondence he told Felix Frankfurter that he was inspired by Frankfurter’s 1927 book decrying the conviction and execution of the Italian immigrants as wrongful convictions (Frankfurter, Reference Frankfurter1927).

Borchard’s Reference Borchard1932 Convicting the Innocent was a continuation of work that he started in 1912, urging American legislatures to enact laws to provide compensation for those who were wrongfully convicted and who could subsequently prove that they were innocent and not at fault for their wrongful conviction. Borchard was a skilled comparative and international law scholar, but he was not a criminal lawyer. His work displayed no interest in the presumption of innocence.

In 1912, while working for the Library of Congress, Borchard prepared a report on compensation for miscarriages of justice in a wide range of European countries. Among his findings were that: “[i]ndemnification both for acquittal on appeal after a conviction, and for detention pending trial followed by acquittal or discharge is provided for in Sweden, Norway, Denmark, Germany, Hungary, Berne, Fribourg, Neuchatel, Basle and Tessin” (Borchard, Reference Borchard1912: 697). He recognized that a number of European jurisdictions provided compensation for a broad range of miscarriages of justice, including not only wrongful convictions but also unwarranted pre-trial detention.

In 1914, Borchard proposed draft American legislation for compensation. It required claimants to prove both their innocence and lack of fault in their own wrongful convictions:

In both the above-mentioned sections, Borchard placed the burden of proof on the applicant for compensation without adverting to the presumption of innocence.

In his efforts to persuade legislators in all American jurisdictions to enact compensation legislation, Borchard deliberately focused on what he called “the most flagrant injustice.” He admitted that “the case of unjust detention pending trial is left aside for the present, in order that we may deal with the much more flagrant injustice of a conviction of an innocent person followed by sentence and imprisonment.” He added: “[i]t is clear that we cannot compensate every acquitted person. In fact, under our lax administration of the criminal law and the possibility of technicalities producing injustice, we know that many morally guilty persons are legally acquitted” (Borchard, Reference Borchard1914: 111). Borchard’s concern about the factually innocent also led to a concern about the factually guilty escaping punishment, again without reference to the presumption of innocence or other due process rights.

Borchard, in comparative law scholar Maximo Langer’s term, “translated” (Langer, Reference Langer2004) more generous European approaches to miscarriages of justice for American audiences. Borchard was concerned with conserving or rationing funds that would be paid to victims of miscarriages of justice in the United States. He also appealed to what James Q. Whitman (Reference Whitman2005) has identified as the United States’ moralistic, populist and punitive approach to criminal justice in ensuring that those who received compensation were seen as morally worthy because they could prove that they were innocent and not in any way at fault for their own wrongful conviction. He was concerned that “the morally guilty” (Borchard, Reference Borchard1914: 111) should not receive compensation even after they were acquitted.

Borchard’s focus on proven factual innocence was successful in both charting a path for American wrongful conviction scholarship and persuading some American legislatures to enact the restrictive compensation legislation he had drafted. In 1913, both California and Wisconsin enacted statutes on the proven innocence model with short six-month limitations for claims, damages limited to pecuniary loss, statutory caps on damages and provisions denying compensation to claimants found to be at fault in any way for their wrongful convictions (Delvac, Reference Delvac2021). These laws were, as one reviewer of Convicting the Innocent noted, “puny” statutes (Hale, Reference Hale1932: 1172). Borchard admitted in his 1932 book, albeit only in a footnote, that only three people had been compensated under the restrictive state compensation laws he championed (Borchard, Reference Borchard1932: 383 fn 52).

Borchard’s proposals for compensation legislation, though debated in Congress in 1912, only resulted in federal compensation legislation in 1938. Borchard assisted in the drafting of the federal law and reassured elected officials that the new law would not result in compensation in “too many cases” (Zalman, Reference Zalman2020: 138). The 1938 federal law required either a court or a pardon to certify that “the claimant did not commit any of the acts with which he was charged” or any other crime. In addition, claimants were ineligible if they “intentionally, or by willful misconduct, or by negligence, contributed to bring about his arrest or conviction” (Pub. Law 1938: ch 266 s.2). As Borchard had recommended in 1914, the federal law capped compensation at $5,000.

The 1938 law still provides the basis for the current federal law. The Justice for All Act, 2004, increased the statutory caps on compensation to $100,000 for each year “unjustly sentenced to death” and $50,000 for every year of imprisonment off death row (28 USC 2513(e) amended by Pub. Law 108–405). In practice, however, the federal law remains “puny” (Hale, Reference Hale1932: 1172), with only 2 of 118 people who have received remedies for wrongful convictions for federal crimes, as recorded in the National Registry, having received compensation under the federal statute. One received just over $137,000 after being wrongly imprisoned for 2.5 years for drug offences and carjacking, and another received almost $552,000 for being wrongfully imprisoned for 12.4 years for drug offences (Gutman, Reference Gutman2021: 238–239). Proven innocence is a popular and durable standard in American law. But it comes with the price of severely rationing justice.

2.4.2 Proven Innocence Migrates from Compensation to Habeas Corpus Law and DNA Access Law

In Chapter 6, a fuller story of how proven innocence has migrated from American compensation laws to DNA access laws and post-conviction relief will be told. The success of this migration can be measured by the fact that the word “actual” or “factual innocence” appears in more than fifty judgments of the United States Supreme Court.

Before being accepted by the courts, factual innocence had academic champions in addition to Borchard. Harvard Professor Paul Bator warned that the liberal Warren Court was promoting through its broad use of federal habeas corpus “an endless repetition of inquiry into the facts and the law” in a futile search for perfection because “no court, can guarantee a result ‘correct’ in an ultimate sense” (Bator, Reference Bator1963: 509–510). Bator, who served in the Reagan administration and afterwards left Harvard for the University of Chicago because of the leftist Critical Legal Studies movement, wanted to restrict federal habeas corpus in 1963 partly because he thought the United States was approaching “the day when Southern justice is like Northern justice, justice for the Negro like justice for the white” (Reference Bator1963: 524). As will be seen in Chapter 6, Professor Bator’s call for deference to the states has been eagerly embraced by the conservative Rehnquist and Roberts Courts. Bator’s scepticism about ever guaranteeing a correct result and his complacency about the functioning of the American criminal justice system were inherited by Justice Scalia, who was a student at Harvard when Professor Bator taught at that school (Kovarsky, Reference Kovarsky2016).

Judge Henry Friendly also sought to restrain federal habeas corpus. He championed a 1969 dissent by Justice Hugo Black that “would require that the convicted defendant raise the kind of constitutional claim that casts some shadow of a doubt on his guilt” (Kaufman v. US 394 US 217, 242 (1969)). This factual innocence requirements, unlike many subsequently imposed by the United States Supreme Court, at least paid attention to reasonable doubt principles by requiring the accused to “show a fair probability that, in light of all the evidence … the trier of the facts would have entertained a reasonable doubt of his guilt” (Friendly, Reference Friendly1970: 160).

The first mention of the phrase “actual innocence” by the United States Supreme Court was in 1986 when Justice Brennan commented that the right to counsel was not “conditional upon actual innocence. The constitutional rights of the criminal defendants are granted to the innocent and the guilty alike” (Kimmelman v. Morrison 477 US 365, 380 (1986)). Justice Brennan used “actual innocence” in a pejorative sense in his defence of due process. In the same year, however, the Court introduced an actual innocence requirement for a successive federal habeas corpus application based on a violation of a constitutional right. Justice Brennan again dissented. He argued that, while a showing of actual innocence satisfied the interests of justice, it should not be the exclusive test (Kuhlman v. Wilson 477 US 436 (1986)). Justice Brennan’s approach was consistent with a broad concern about miscarriages of justice that included, but were not limited to, proven innocence.

In 1992, the Supreme Court denied a successive habeas corpus petition on the basis that the petitioner had not by clear and convincing evidence shown actual innocence as a precondition or gateway with respect to mitigation evidence relevant to the death penalty (Sawyer v. Whitley 505 US 333 1992). The concept of innocence that had originally been related to reasonable doubt about guilt now required proof of innocence by the convicted person on the basis of clear and convincing evidence. Justice Stevens concurred but argued that actual innocence should not be the exclusive basis for a miscarriage of justice exception because “justice is more than guilt or innocence” (Sawyer v. Whitley 505 US 333, 361). This followed Justice Brennan’s inclusive approach to miscarriages of justice.

Fast forward to today, and a conservative Court is cutting back on using proven innocence as a means to circumvent restrictions on habeas corpus. In Jones v. Hendrix, 143 S.Ct. 1857 (2023), Justices Sotomayor and Kagan stressed that the majority was denying post-conviction relief to a “prisoner who is actually innocent for conduct that Congress did not criminalize.” In her dissent, Justice Jackson alluded to a “non-frivolous argument that the Constitution’s protection against ‘cruel and unusual punishment’ prohibits the incarceration of innocent individuals” while also inviting Congress to increase the ability of convicted persons to make “innocence claims.” Liberal Justices have gone from resisting proven innocence as an exclusive test of injustice to defending proven innocence as an exception to the Court’s increasingly restrictive approach to habeas corpus relief. Again, innocence is being used as a means to ration access to post-conviction relief.

2.4.3 The Popularity of Proven Innocence

A focus on factual or proven innocence represents what Michael Naughton has defined as “lay” understandings of justice (Naughton, Reference Naughton2013). It is thus understandable why both elected legislators and the elected executive have been responsive to clear cases of proven innocence. Julia Carrano and Marvin Zalman have recognized that justice is rationed when law reformers focus on proven factual innocence because such a category “includes the fewest numbers of defendants and carries the greatest moral weight. Moreover, it signals a particularly great public safety concern – the real perpetrator has gone free and is capable of committing additional crimes” (Carrano and Zalman, Reference Carrano and Zalman2014: 12). Jon Gould, (Reference Gould and Zalman2014a: 5–6) similarly warned innocence reformers “need not ‘sell’ changes on the basis of civil rights or due process.” He warned “we simply are not going to achieve the reforms necessary to prevent erroneous convictions if the arguments are coached in the language of defendant’s rights or if reformers are characterized as anti-police or prosecutors.” This focus on factual innocence and the false idea that a guilty person goes free in every wrongful conviction case helps explain why retentionist and traditionally conservative states like Texas and North Carolina have been at the forefront of bipartisan wrongful conviction reforms, topics to be discussed in Chapter 6.

2.5.1 Nelson v. Colorado

In 2017, the United States Supreme Court ruled that a Colorado statute that required applicants to prove their innocence by clear and convincing evidence to recover court fees collected upon their conviction violated due process. The Court’s judgment rejected the formalistic distinction that the presumption of innocence did not apply because the procedure to recover court fees was characterized as a “discrete civil proceeding.” Justice Ginsburg stressed that “once convictions were erased, the presumption of their innocence was restored” (Nelson v. Colorado, 581 U.S. 128, 130, 135). After the presumption of innocence was treated as a fundamental principle and not simply a way of allocating burdens of proof and risks of error in a criminal trial. Once a conviction was overturned, society should treat those it previously convicted as innocent and equal members of society. Although it was principled, the 2017 decision was also hypocritical given that the present United States Supreme Court has imposed proven innocence requirements in its attempt to restrict habeas corpus relief in Federal Courts (Re Davis, 130 S.Ct. 1 2009; McQuiggan v. Perkins, 569 U.S. 383, 2015).

In his concurrence, Justice Alito endorsed the formalistic distinction between criminal trial and other proceedings when he stressed that, because a request for compensation was not a criminal trial, “the burden of proof can be placed on the petitioner” (2017 at 145). He argued that this formalistic approach was “best justified by history and tradition” (2017 at 146) without noting that the history started with Edwin Borchard’s successful advocacy for a proven innocence standard.

This 2017 decision, however, appears to be a blip of principle. Following Justice Alito’s formalistic approach, which limits the presumption of innocence to criminal trials, there have not been successful challenges to compensation statutes that require previously convicted persons to prove their innocence. It has also not caused the Court to back away from proven innocence standards used to restrict access to federal habeas corpus. American courts were more concerned with the presumption of innocence when a previously convicted person tried to collect court fees than when they sought their liberty.

2.5.2 The European Convention on Human Rights

In a 2011 compensation case of Adams, Lady Hale argued that an approach to compensation that required proven innocence was in tension with the “golden thread” of the presumption of innocence because the accused “does not have to prove his innocence at his [criminal] trial” and “it seems wrong in principle that he should be required to prove his innocence now” to obtain compensation (R (Adams) v. Secretary of State for Justice [2011] UKSC 18 at para 116). Similar to Justice Ginsburg’s opinion in Nelson v. Colorado, this was a broad principled approach that seemed prepared to apply the presumption of innocence outside of the criminal trial context.

By 2019, however, Lady Hale joined a majority that refused to hold that an innocence requirement added by Parliament in 2014 to English compensation legislation was inconsistent with the presumption of innocence. Her ruling seemed to be based more on a realpolitik and ultimately correct prediction that it was unlikely that the European Court of Human Rights (ECtHR) would find England’s new law inconsistent with the presumption of innocence as guaranteed in the European Convention on Human Rights (R (on application of Hallam) v. Secretary of State 2019 UKSC 2 at para 82).

Lord Hughes took a formalistic approach similar to that of Justice Alito by categorically declaring that a person applying for compensation “is in no danger of conviction or punishment. Nor is he in any danger of any official body treating him as if he were still convicted or liable to punishment” (2019 at para 124). This was a thin and impoverished understanding of the presumption of innocence that mirrored boiler plate statements made by the Secretary of State that the denial of compensation was not “intended to undermine, qualify or cast doubt upon the decision to quash [the] conviction” and that the applicant “is presumed to be and remains innocent brought against him” (2019 at para 15). In contrast, Hoyle and Tilt (Reference Hoyle and Tilt2020: 49), after interviews with those involved in compensation claims denied on such basis, have convincingly concluded that “denial of compensation was experienced by clients of the lawyers we interviewed as an indication that they were not sufficiently innocent to meet the threshold for compensation. Applicants felt confused and insulted by these decisions.” They also point out that the decision of the elected executive, the Secretary of State, to deny compensation can only be challenged through the deferential administrative law standards of judicial review. Proven innocence allows standards of civil and administrative justice that are much less demanding on the state than criminal justice standards to prevail.

Only two judges of the UK Supreme Court in dissent would have declared the proven innocence requirement in the 2014 compensation law to be incompatible with the presumption of innocence. Lord Reed warned that it was improper for the executive to rely on suspicions about the previously convicted person’s guilt and that doing so “casts doubt on the innocence of the person in question and undermines the acquittal” (R (on application of Hallam) v. Secretary of State 2019 UKSC 2: para 184). Lord Kerr warned: “establishing innocence as a positive fact can be an impossible task” (R (on application of Hallam) v. Secretary of State 2019 UKSC: para 203), especially given that the courts that overturned the conviction do not make determinations of innocence.

On appeal, a majority of the Grand Chamber of the ECtHR decided that proven innocence requirements did not violate the presumption of innocence because it did not impute criminal guilt “in the eyes of the law” to applicants who were unable to prove beyond a reasonable doubt that they did not commit the offence (Nealon v. United Kingdom, App. Nos. 32483/19 & 35049/19, 178–181 (June 11, 2024)). This was a formalistic approach that stressed that compensation was an administrative or civil matter apart from the criminal law. It was also unconvincing in its uncritical acceptance of statements by English executives that, in denying compensation to the formerly convicted applicants, including Victor Nealon, who had been acquitted by a DNA exclusion, they were somehow not casting doubt on their innocence. The reality, however, is that exonerees and a suspicious public understandably experience such denials of compensation as a denial of their innocence (Hoyle and Tilt, Reference Hoyle and Tilt2020). The Grand Chamber also equated compensation for miscarriages of justice with compensation for crime victims (Nealon v. United Kingdom at para 153, 165), ignoring that the state bears more direct responsibility for wrongful convictions than a failure to prevent crime and focusing on issues of factual as opposed to legal guilt.

Five judges in dissent took a less formalistic approach that went “beyond the mere wording of the decision casting doubt on the person’s innocence” (Nealon v. United Kingdom at para 5). They concluded that the 2014 English law violated the presumption of innocence because it embraced a “presumption that the applicants committed the criminal offence, a presumption which can only be rebutted if they show beyond reasonable doubt that they did not commit it” (Nealon v. United Kingdom at para 5). The dissenters also pointed out that the new English approach was much less generous than most European states, which “provide for strict State liability in the event of miscarriages of justice” including in some cases of pre-trial detention that end in an acquittal (Nealon v. United Kingdom at para 8). Even the majority of the Court noted that most European states rejected the English (and American) proven innocence approach (Nealon v. United Kingdom: paras 95–96) but was willing to accept this English exceptionalism under the European Convention on Human Rights on the basis that “criminal justice systems varied greatly according to each Contracting State and its legal traditions” (Nealon v. United Kingdom: para 97). The Court was perhaps reluctant to rule against another English law after having had England largely ignore its decisions that prisoners should be able to vote and when some in England were arguing that it should follow Brexit by also renouncing the jurisdiction of the European Court of Human Rights (Hodgson and Roach, Reference Hodgson and Roach2017).

The effect of the ECtHR’s ruling, however, went well beyond England. Now, all European states could, if they want, require those who had been wrongfully convicted to prove their innocence beyond a reasonable doubt in order to obtain compensation. The generous standards of countries such as France, Germany and Italy that Borchard noted compensated victims of both wrongful detention and wrongful convictions could be replaced with his populist and moralistic concept of proven innocence without violating the European Convention on Human Rights. This underlines the fragility of more generous and principled approaches to miscarriages of justice.

Unless one accepts a formalistic and categorical divide between criminal justice matters on the one hand and administrative and civil justice matters on the other hand (as do Justice Alito in Nelson v. Colorado and the ECtHR in Hallam v. United Kingdom), proven innocence requirements are at odds with the presumption of innocence. Justice Ginsburg in Nelson v. Colorado, Lady Hale in Adams and the dissenters in Hallam have all recognized that laws that require proof of innocence allow a person to be treated as guilty even if there is a reasonable doubt about their guilt. Alas, there is not much popular support for such an approach, given public opinion polls that suggest that the public views the convictions of innocent and the acquittal of the guilty as equal injustices to be avoided.

2.5.3 The Presumption of Innocence and the Blackstonian Ratio Under Siege

One study based on global public opinion polls and experiments with students shows decreasing support for the principle that it is better that ten guilty people go free than one innocent person be convicted, articulated by the English lawyer Blackstone in 1770. Blackstone’s ratio has origins in the Bible, theorists such as Aristotle and Maimonides, Roman law and the work of earlier English lawyers (Epps, Reference Epps2015: 1070–1077). A 2005 General Social Survey in China found that 44.4 per cent of respondents believed that wrongful convictions were worse than wrongful acquittals, but 42.2 per cent believed that wrongful acquittals were worse with the rest saying it was too difficult to decide (Xiong, Greenleaf and Goldschmidt, Reference Xiong, Richard and Jona2017: 18). This view is not restricted to China because studies showing that 51.4 per cent of respondents in a British survey viewed wrongful convictions as worse, but 48.6 per cent viewed wrongful acquittals as worse (Reference Xiong, Richard and Jona2017: 17). A 2006 international survey from thirty-eight countries found 62 per cent viewed wrongful convictions as the worst injustice, but 37.9 per cent of respondents disagreed (Reference Xiong, Richard and Jona2017: 18). The researchers used these findings to propose that the standard of proof for “terrorists and mass murderers” be lowered to a “clear-and-convincing standard, or even a preponderance -of-the-evidence standard” in order to gain greater “public acceptance” (Reference Xiong, Richard and Jona2017: 23). In subsequent work, the lead researcher of the above-mentioned study, Chinese law professor Moulin Xiong, has argued that these public attitudes justify greater stress on crime control. Professor Xiong cites in support American law professors such as Paul Cassell (Reference Cassell2018) and Daniel Epps (Reference Epps2015), who similarly argue that the Blackstonian ratio devalues the harms of wrongful acquittals (Xiong, Reference Xiong2022).

American law professors Brandon Garrett and Gregory Mitchell (Reference Garrett and Mitchell2023: 721) have, on the basis of their own public surveys, concluded that 61 per cent of American respondents are equally as concerned about wrongful acquittals as wrongful convictions, with 25 per cent viewing wrongful convictions to be more harmful and 14 per cent viewing wrongful acquittals as more harmful. These attitudes were consistent regardless of the respondent’s willingness to convict or whether they were aligned with the Democratic or Republican Parties. These findings suggest that there is likely to be increasing public resistance to the presumption of innocence, the reasonable doubt standard and preferences for avoiding wrongful convictions over wrongful acquittals. This creates an environment in which justice for the accused may be more severely rationed.

One way to respond to the above-mentioned public concerns about wrongful acquittals that is less obvious and jarring than Professor Xiong’s call for using lower standards of proof for serious crime is to demand proven innocence in compensation and post-conviction proceedings. A 2006 attempt by the Blair government to deny appeals from convictions from those who were not innocent was withdrawn after widespread protests by lawyers and even the CCRC. That said, the United States has embraced proven innocence as a means to ration post-conviction relief, and China also de facto uses a standard of proven or obvious innocence in remedying wrongful convictions. Countries, taking cues from the United States and England, that have been prepared to give the accused the benefit of a reasonable doubt may in the future be more inclined to require proven or obvious innocence. Indeed, even in cases where proven innocence can be proved, these public opinion survey suggests that many will be just as concerned about wrongful acquittals and a significant minority may be more concerned about wrongful acquittals than wrongful convictions.

2.5.4 The Dangers of the Certainty Required for Proven Innocence

The demand for proven innocence often exudes a certainty that is rarely found in the real world or in science. In a thoughtful article, Keith Findley has observed that “innocence has no real legal meaning in most jurisdictions” (Findley, Reference Findley2011b: 1160). Although Professor Findley co-founded and headed the University of Wisconsin’s Innocence Project, he wisely warns that it is a mistake to demand certainty, something that the criminal justice system does not even require for a conviction. He thus argues that while innocence is important, it should expand beyond the “closely guarded” list where DNA evidence is interpreted to produce clear-cut innocence. He warned that compensation statutes and other proven innocence requirements can create two classes of innocent people: the “super innocent” who are lucky enough to be able to prove their innocence and the not so innocent (Reference Findley2011b: 1208). All of this suggests that courts and legislators should think twice before adopting a proven innocence standard even if the public is as concerned about wrongful acquittals as wrongful convictions.