Proving Unlawful Discrimination in Capital Cases: In Quest of an Adequate Standard of Proof

In spite of some early judicial, political and scholarly discussions, as well as more recent scientific explorations of the topic, problems and concerns with proving discrimination in individual capital cases continue to be among the most debatable issues in human rights and criminal justice. In general, domestic courts (in particular US courts) seem to remain relatively perfunctory and hostile to individual discrimination challenges in capital trials. They normally require capital defendants alleging discrimination to prove something which is virtually impossible to prove. On the other hand, numerous capital defence attorneys, legal commentators and even some of the trial judges themselves lay strictures on the existing judicial approach which almost routinely rejects discrimination claims in capital cases. They contend that appropriate modifications in current legislative arrangements and mechanical adjudication policy and practice are urgent and indispensable for more equitable resolutions and for a truly even-handed criminal justice system. In particular, there are concerns regarding the adequate distribution of the burden of proof between the litigants. Moreover, no clear or uniform approach to this conundrum can be identified in the international jurisprudence. This article seeks to provide some definite answers to open and conceptual questions posed in an attempt to legally define ‘the minimum core content’ of the evidentiary standard – as implicitly contained in the relevant international human rights treaties’ provisions – to be applied in capital sentencing discrimination cases. Additionally, part of this same standard of proof can also qualify as a general principle of international law, particularly in relation to impartial, unbiased and non-discriminatory approaches and decisionmaking by the judges and jurors involved in complex capital cases.


I. Introduction
Discrimination in the administration of the death penalty constitutes a substantial segment of miscarriages of criminal justice in death penalty jurisdictions. Given the wide range of discretion assigned to decisionmakers involved in capital cases, pervasive bias, prejudice and stereotypes related to race, national or ethnic origin, gender, religion or belief, socio-economic status, place of origin, sexual orientation, disability, level of education and similar characteristics (grounds of discrimination) of defendants/victims involved in capital crimes can play a role in decision-making processes in capital trials. Some of these decisions include a prosecutor's decision to file an indictment and seek a death penalty for an alleged perpetrator or offer a plea bargain, and the ' careful' selection of jury members, or the fact-finder's final decision concerning a capital defendant's culpability and punishment.
Obviously, it is extremely hard to prove such hidden and covert preconceptions and stereotypical attitudes which often lead to discriminatory decisions in capital cases, unless they are articulated in some way. In some instances, decision makers in capital cases may not even realise that they are applying stereotypes in their decisions or acting discriminatorily because the ideas are so deeply ingrained or institutionalised within a criminal justice system. Moreover, prosecutors, jurors and judges will always be astute enough to think of ingenious rejoinders to rebut claims of alleged disparities in the application of the death penalty and it will be impossible to elicit statements admitting or indicating when a discriminatory decision was adopted in a particular capital sentencing process. Hence, without tangible evidence of biased treatment in their particular case it is extremely difficult, if not impossible, for those claiming to have been subjected to discrimination during the capital sentencing process to substantiate their claims and to prevail upon the judge hearing their case.
Providing sufficient evidence of discrimination in the capital sentencing process is a central issue of civil procedure when it is claimed that the death penalty was sought and/or imposed on a defendant in a discriminatory manner. Against that backdrop, the question arises as to what and how much evidence must be provided by victims of discrimination in death penalty cases to establish a prima facie case of discrimination so as to successfully meet the requirement for shifting the burden of proof to the respondent (state). This article addresses these key questions in order to develop an adequate and contemporary standard of proof that courts and other competent adjudicative authorities should apply as far as the issue of proving discrimination in capital sentencing processes is concerned.
At the outset, it is fair to say that this article attempts to address the research problem primarily from international legal perspectives and thus complements and upgrades the existing body of American scholarly research and literature in this area which is, for the most part, concerned with domestic legal and judicial considerations. These predominantly federal-and states-oriented analyses and evaluations usually overlook a simple but very intriguing fact that there is an absence of a clear, uniform and well-defined approach to the issue that is currently witnessed at the international level and bespeaks a great unease towards identifying an appropriate standard of proof within domestic death penalty legal systems. While some specialised international human rights treaty bodies, such as the Committee for the Elimination of Racial Discrimination and the Committee on the Elimination of Discrimination Against Women, chose to espouse the modern standard of proof which is reminiscent of the concept codified in the advanced European Union (EU) non-discrimination law -according to which, the rule on shifting the burden of proof lies at the heart of evidentiary procedure in discrimination cases and aims at somewhat easing the claimant's burden of proof in civil proceedings -the Inter-American Commission on Human Rights seems to opt for a more conservative, reserved and prudent, if not retrograde, approach. 1 The Inter-American Commission's approach requires a significantly higher standard of proof which asks the claimant to present direct, abundant and clear-cut evidence of discrimination (as evident from the Commission's findings in Celestine and Andrews) 2 and is thus much closer to the US Supreme Court's reasoning in the famous McCleskey case demanding a discriminatory purpose on the part of decision makers in capital sentencing be proven, 3 thereby calling into question its compliance with international non-discrimination standards as correctly observed by some authors and commentators. 4 By adding an international legal dimension to the subject, it is argued in the present article that the standard of proof to be employed in civil procedures dealing with discrimination in capital sentencing cases constitutes, in part, a general principle of law recognized by (civilized) nations and therefore, at least in this respect, should be one and the same regardless of the jurisdiction or criminal justice system and irrespective of the form or type of discrimination alleged. Moreover, these same rules of proving discrimination in capital cases should provide for an equitable and balanced distribution of the burden of proof and also enable specific evidence -such as relevant statistics -to serve as a sufficient proof of discriminatory effect; they would thus constitute an instrument which can trigger a shift in the burden of proof. This position is primarily based on the extant international human rights jurisprudence evolved through the interpretation of relevant international human rights treaties' provisions by the Human Rights Committee, the Committee on the Elimination of Racial Discrimination, the Committee on the Elimination of Discrimination Against Women, the European Court of Human Rights, the Court of Justice of the European Union and also the European Committee of Social Rights.
Furthermore, it is argued that, in view of contemporary insidious bias and discriminatory attitudes of the decision-makers involved in capital cases, such standard of proof should be much closer to the one applied in civil law cases of discrimination. This assertion can find its foothold in the existing legal scholarship, relevant international and domestic case law and jurisprudence, and current evolutionary trends in discrimination law. The central thesis is that especially the rule on the shared burden of proof -as defined and applied in the context of advanced EU anti-discrimination law -should serve as a good example of a minimal, decent and just standard of proof to be followed in proving discrimination in capital cases in order to amend the status quo and remedy the existing state of affairs. This underlying thesis elaborates on several relevant starting-points and criticisms and is predicated on some crucial international legal arguments and remarks which are outlined and discussed below.
Following a critical appraisal and in-depth analysis of relevant international and domestic (case) laws, it is argued that -against the backdrop of contemporary practices and forms of discrimination which normally are and remain subtle and insidious -the methods of proving discriminatory decision-making procedures in criminal trials in general, and in capital cases in particular, call for an appropriate transposition of the evidentiary machinery introduced and established by the EU non-discrimination directives if the international fair trial and equal treatment standards are to be utilised and enforced to their full potential. Subsequently this article attempts to formulate some critical remarks and give some reasoned arguments why the standard of proof worked out and pursued by some domestic courts -such as those from US death penalty jurisdictions -and international quasi-judicial bodies, such as Inter-American Commission on Human Rights, ought to be seen as old-fashioned and inadequate in light of contemporary discrimination claims.

II. International, European and Inter-American legal approaches to discrimination in death penalty cases A. Establishing Direct Discrimination in Capital Cases
If, for example, a certain legislator enacts a law which sanctions homo-and bi-sexuality with the death penalty, whereas heterosexual relations are not punishable at all, this may result in many more homo-and bi-sexuals being sentenced to death and executed than those having preference for the opposite sex. However, nowadays, cases and situations where discrimination is manifested or expressed in such a direct and easily identifiable way are extremely rare. This holds true also for capital sentencing processes. Thus, discrimination in capital sentencing decision-making processes is in many cases subliminal and invidious. As a result, proving direct discrimination in capital cases is usually a very demanding task, although, by definition, the difference in treatment is ' overtly' based on one or more characteristics of the capital defendant and/or of the victim of a capital offence. The fact is that decision-makers in capital cases will not always declare their differential treatment between persons of various backgrounds, nor will they reveal their reason(s) for such treatment.
For example, an Afro-American capital defendant may be sentenced to death with a simple explanation that 'specific circumstances in which the capital crime was committed and the principle of justice' in their particular case call for the ultimate sentence and punishment. However, there are several precedents where white defendants were sentenced to life imprisonment for the same capital offence in similar circumstances and with a similar criminal history. In this situation, the capital defendant who would believe that the death penalty was imposed on them in a biased and discriminatory manner may find it difficult to prove that they were directly discriminated against because of their race or ethnic origin.
This section considers and analyses several issues which are of key importance to the interpretation, realisation and effectiveness of statutory provisions on proving direct discrimination in the context of the death penalty application. These include the question of causation and identifying a proper comparator in establishing direct discrimination, the relevance of stereotyped and prejudiced views as well as discriminatory intent and motive on the part of decision makers, and the meaning of the notion of less favourable treatment in the capital sentencing context.

B. Identifying a Suitable Comparator and Causation
One of the core conditions required by discrimination law (including relevant EU non-discrimination directives) is that the complainant claiming the existence of direct discrimination is able to show that they have been treated less favourably than another person is, has been or would be treated in a comparable situation and that such difference in treatment was due to prohibited characteristic(s). 5 This means that two persons or categories of persons who are in comparable situations and whose factual and legal circumstances uncover no essential difference should be treated in the same way without regard to personal traits, such as race, national or ethnic origin, gender or socio-economic status. At the same time, this general principle of equal treatment, enshrined in all core international human rights instruments, requires that such treatment is not identical in situations that are different. In the context of the death penalty application, this implies that prosecutors, jury members and judges as decision-makers must not take into consideration legally irrelevant characteristics when making decisions on whom to sentence to death or not. Their decisions must be fair and rational, based solely on legally relevant factors.
Indeed, locating an actual comparator in each and every case of direct discrimination can be problematic. For instance, in many death penalty jurisdictions there are well-documented patterns of gender segregation in the application of capital punishment. In situations where death row is dominated by persons of a particular gender, in practice it may be very difficult to identify an appropriate actual comparator. Here, the requirement for a comparator does not seem to be helpful in identifying direct discrimination. Some commentators therefore argue that in some instances of direct discrimination it is not necessary to find a comparator since in some situations direct discrimination will be particularly manifest and self-evident. 6 The importance that is ascribed to the comparator identification in direct discrimination cases varies considerably from one legal system to another. While some domestic legislation places rather strong emphasis on identifying an appropriate comparator, 7 others prefer enquiring into the question of causation, that is, whether the reason for less favourable treatment was discriminatory; although the identification of a comparator in circumstances similar to those of the complainant is not completely omitted, it becomes one of the possible ways to establish causation. 8 Furthermore, the difference can be noticed in the order in which domestic courts and equality bodies apply the comparator requirement in direct discrimination cases. 9 It will not always be possible to choose the ' correct' comparator. The difficulties with a comparator requirement may emerge especially in cases dealing with discrimination on grounds of sex, gender reassignment, sexual orientation and age. Let us assume that a person who underwent gender reassignment (from male to female) is sentenced to death because she is, in spite of changing her sex, still treated as a male in a certain legal system applying the death penalty. Indeed, if her new gender was recognised as legally relevant she would be able to escape the death penalty. The person claims that the capital punishment imposed on her was a result of gender discrimination in her capital sentencing process. The question which arises in such a case is what is an appropriate comparator in establishing less favourable treatment of the person who underwent gender reassignment. Here, the classic model of a straightforward comparison between men and women can hardly be applied, which is why more than one comparator could be ' correct'. One may suggest, for instance, that the treatment of the person claiming that the death penalty was imposed on her in a discriminatory manner should be compared with that of a man or a woman who did not undergo gender reassignment. The result of such a comparison is the same and shows that the claimant had been disadvantaged. But is a person who has not undergone gender reassignment really in a similar situation to the person claiming gender discrimination? Such an approach can also be understood as suggesting that the complainant herself, prior to undergoing gender reassignment, could serve as a comparator. Another option in this case is to assert that the comparator requirement can and should be dispensed with.
Similar problems in finding an appropriate comparator may arise in cases of discrimination on grounds of sexual orientation. Here again, it seems to be questionable in cases involving same-sex couples and whether 5  those couples are really in a comparable situation to either married or unmarried opposite-sex couples. The issue of preferential or disadvantaged treatment on the basis of the perpetrator's sexual orientation may also be invoked in capital cases involving defendants who are gay or lesbian.
Two methods have been employed by the European courts and equal treatment bodies to rectify the shortcomings of a comparator requirement in cases in which it is difficult to define the precise approach to be pursued when selecting an appropriate comparator, especially in situations where more than one comparator can be potentially ' correct' and when -as demonstrated by the above cases dealing with gender reassignment discrimination -it has turned out that actually no comparator will be completely appropriate. Mention has already been made of one of the possible techniques to overcome such difficulties in direct discrimination cases -that is, the complete departure from the comparator requirement. The other useful alternative is the possibility of invoking a hypothetical comparator.
The EU legislature accepted this reality and included the hypothetical comparator aspect in the standard definition of direct discrimination contained in the EU non-discrimination directives: ' direct discrimination shall be taken to occur where one person is treated less favourably than another is, has been or would be treated in a comparable situation'. 10 According to this definition, one needs to make a comparison between the treatment of the victim of discrimination on the one hand, and of a comparator on the other hand. The comparator may be actual as reflected in the wording 'is treated' and 'has been treated' or may be hypothetical as captured in the text 'would be treated', but the relevant circumstances in the one case must be comparable to those in the other. Most EU member states have incorporated such express reference to hypothetical comparators into their domestic non-discrimination legislation in the process of implementation and transposition of EU anti-discrimination directives into their national legal systems.
In a particular case, where an actual comparator capable of constituting the statutory comparator cannot be found, a hypothetical comparator can be used instead. Indeed, most challenging in applying a hypothetical comparator is the necessity to objectively assess a hypothetical situation, that is, a situation that did not actually exist. In this regard, the interpretation of 'how a hypothetical comparator would have been treated' 11 by taking into consideration the broader context of a given case may be of particular assistance. In other words, the actual treatment of persons who are not in a sufficiently similar situation to that of a complainant, and thus cannot serve as an actual comparator, may offer important evidence for considering how a hypothetical comparator would have been treated.
In addition, a court should pay due regard to any other existing evidence from which discrimination may be inferred, such as racially offensive statements, inscriptions, discriminatory questions, or sexist comments made during the course of a hearing or adopting the jury's decision on the defendant's culpability and application of capital punishment. Further, inferences of discrimination may also be drawn, for example, where a capital jury's function has not been exercised in compliance with the court's normal procedural rules and where there is some evidence of hostile attitudes towards minority group members that the capital defendant belongs or is perceived to belong to within the jury panel. Such pieces of evidence can be very relevant since they can replace the need for an actual comparator by comparing a person claiming discriminatory application of the death penalty with a hypothetical person with different personal characteristics. This will enable a court to consider whether such hypothetical comparators would have been treated differently from the capital defendant complaining of a discriminatory application of capital punishment in their case.

C. Sorting Out Stereotypes-and Prejudice-Based Decisions in Capital Cases
In order to better understand the crux of this appraisal we must first identify the role that stereotypes as collective social representations and social psychological constructs play in capital sentencing decision-making processes. Stereotypes and prejudice that members of different racial, ethnic and religious groups hold towards each other are the source of all the evil caused by discrimination. This is also true of discriminatory treatment in capital sentencing processes. Needless to say, stereotypes and prejudice that main legal actors involved in the death penalty cases (prosecutors, jurors, trial judges and defence counsel) 12 harbour may influence their decisions and acts. How strong this impact is depends, inter alia, on the nature, level and amount of bias that the decision makers are infected with and on their (in)ability to overcome such stereotypes and prejudice when making decisions in the capital sentencing process.
Stereotypes often serve as explanations or justifications for direct discrimination. However, some of the stereotyped preferences that discriminators hold can be shrouded and thus will not necessarily be supported by firm contextual evidence. One should be aware of the fact that contemporary forms of stereotyping and prejudice have taken subtler and more intractable characteristics than in the past. With time, the incidents of intentional and blatant forms of discrimination became socially and legally unacceptable and thus less likely to occur; however, the domination of one (usually majority) group over another (usually minority) group has been maintained and justified, despite explicit egalitarian beliefs, through prejudiced and discriminatory conduct that can most frequently be justified as unintentional and which takes place in a way that appears to be justifiable. This reality is also reflected in the modern criminal justice system, including in capital sentencing cases. 13 Several empirical studies and experimental works, including that carried out by Mona Lynch and Craig Haney, 14 clearly suggest that racial bias often comes into play in death penalty decision-making processes, most probably to the greatest extent in discharging capital juries' tasks. The issue may arise already in the capital jury selection process which allows exclusively for ' death-qualified' jurors and excludes all other possible candidates (especially fervent death penalty opponents). This selection process almost always results in all-white or predominantly white, male, older and more conservative (in the religious and political sense) capital jury members who are more likely to hold wide-spread and persistent cultural stereotypes towards minority groups' members (for example, the stereotype that African-Americans are dangerous, aggressive and genetically predestined and inclined to commit serious criminal offences). Another interesting example of influencing the capital jury's decision-making process by making insinuations of certain most common stereotypes and prejudices in relation to capital defendant and/or victim of a capital crime occurs when the capital prosecutor refers or appeals to (either explicitly or implicitly), for example, racial or gender bias in their argument or other conduct. This is the case in many instances of capital trials as far as the US death penalty system is concerned. 15 Indeed, negative stereotypes and prejudices against minority groups and their members, which are particularly subtle and resistant to change, may (and, at least in the US death penalty jurisdictions, in certain cases did) strongly influence the outcome of a capital sentencing process. Many empirical studies and experiments done by American scholars have confirmed that the circumstances in which laypersons acting as capital jurors make their decisions render it possible for them to rely on stereotypes and prejudice. 16 However, in order to establish a prima facie case of discrimination in the application of the death penalty, a person affected by the arguably discriminatory death sentence will need to show some connection between prejudiced beliefs and stereotyped thinking of capital prosecutors, jury members and/or trial judges and their decision on the defendant's culpability, death-worthiness and ultimate punishment. This is certainly not an easy job, especially not for someone who is destitute and lacks legal expertise and other relevant resources. In addition, some existing domestic legal models of the non-discrimination clause, such as the Fifth and Fourteenth Amendments of the US Constitution, 17 turned out to be old-fashioned and inadequate since they will hardly ever be applicable to the incidents of contemporary forms of insidious and sometimes also unconscious biased behaviours that are neither expressed in a clear and overt manner nor based on any apparent discriminatory motive. 13 Stereotypes and prejudicial attitudes become especially problematic where those responsible for key decisions have broad discretion, as in the decisions to seek or impose the death penalty. See also, in this regard, the Chief Justice Barkett's criticism concerning the problem of proving unconscious and invidious racism on the part of decision-makers in capital cases discussed in  187-190. 17 These non-discrimination provisions require the plaintiff to prove the existence of discriminatory motive or intent (as a subjective state of mind) on the part of the respondent (alleged perpetrator) to successfully establish direct discrimination in their particular case.
As it is very unlikely that people will advertise their stereotypes and prejudices, and in fact they may not even be aware of them, proving discrimination must allow for inference rather than direct evidence, including in the death sentencing context. 18 This entails that once the first three of the mentioned prerequisites (less favourable treatment of one person than that of a comparable person) are clearly shown by the victim of discrimination, the judges will have to ask the alleged discriminator for an adequate explanation which must be unrelated to personal trait(s) of the complainant. If the respondent is unable to provide a satisfactory explanation, it is quite legitimate to infer that the less favourable treatment was on the prohibited ground(s) of discrimination. And if the difference is on one or more such grounds, the reasons or motives behind it cannot be seen as relevant. 19 Non-discrimination safeguards embedded in or relevant to criminal law procedures require prosecutors, jurors and judges to treat each person as an individual and not as a member of a particular group. Therefore, one may safely say that the capital defendant should not be assumed to hold the characteristics which the prosecutor, juror or trial judge associates with a group, whether or not most members of the group do have such characteristics. Even if, for example, it is still a widely held stereotype that most Afro-American men are lawless, more violent, dangerous and criminally inclined than white men, it must not be assumed that an Afro-American who has found himself in a death penalty criminal process is violent, dangerous and inclined to commit crimes and therefore must be punished and sentenced more severely than a white capital defendant who committed the same or a very similar kind of crime and has largely the same criminal record. Such application of a stereotype by a capital prosecutor, juror or trial judge which gives rise to a less favourable treatment of the members of minority groups must be deemed legally impermissible and unjustified. Legislation prohibiting discrimination in the criminal justice system requires that each and every capital defendant is treated as an individual and not assumed to be like other members of the group to which they belong. 20 Most commonly, the evidence of stereotyping and prejudice on the part of a capital prosecutor, jury member or trial judge will not be available, but where such evidence is present (as this was the case, for example, in Andrews v United States), 21 this may certainly contribute to establishing the existence of direct discrimination in the capital sentencing process. On the other hand, if such evidence is absent, or there is proof of a non-prejudicial motive, this should not be considered an obstacle to a finding of direct discrimination in a death penalty case. 22

D. Identifying Bias and Prejudice on the Part of Decision-Makers in Capital Cases
Sometimes it will be possible to prevent the adoption of biased decision(s) in a given capital case by a timely finding that, in fact, one or more key decision-makers are partial or prejudiced against a capital defendant or victim of a capital crime or against both of them. However, to come to such a finding, a trial judge will need to seriously approach and carefully examine any potential danger or suspicion of bias on the part of the decision-maker(s) and if the existence of such a bias is established they will be expected to undertake all necessary steps to eliminate any peril of biased decision making and to ensure a fair, just and impartial criminal process throughout the capital trial. 23 International fair trial standards also include the requirement that the tribunal deciding in the death penalty case is impartial and affords a capital defendant equal protection of the law, without discrimination of any kind. In criminal justice systems that employ a jury system, these requirements apply to professional 18 In McCleskey Justice Blackmun criticised the analysis of the equal protection clause made by the majority of judges for not following its ordinary three-fold test for establishing a prima facie case of discrimination as defined in Castaneda v Partida, Castaneda v Partida, 430 U.S. 482 (1977). Under this test, a defendant needs to show that 1) he or she is a member of a distinct class; 2) there is a substantial degree of differential treatment; and 3) the allegedly discriminatory procedure is subject to abuse or is not racially neutral. When the defendant establishes a prima facie case of discrimination under Castaneda test, the burden of proof shifts from the defendant to the prosecution to rebut the allegation. and lay judges as they do to jurors. 24 Under international and European human rights law both the subjective and objective test based on 'reasonableness and the appearance of impartiality' shall be employed when considering the issue of judge and juror impartiality in criminal cases, including in capital sentencing. 25 Such an approach can be inferred also from relevant case law developed by the international human rights treaty bodies. For instance, the Committee for the Elimination of Racial Discrimination held in Narrainen v Norway that a reasonable suspicion of bias is sufficient for juror disqualification by stating that 'it is incumbent upon national judicial authorities to investigate the issue and to disqualify the juror if there is a suspicion that the juror might be biased'. 26 The European human rights system strictly follows the objective test as well, complemented by the subjective test. In Piersack v Belgium the European Commission of Human Rights stated that: Whilst impartiality normally denotes absence of prejudice or bias, its existence or otherwise can, notably under Article 6 § 1 (art. 6-1) of the Convention, be tested in various ways. A distinction can be drawn in this context between a subjective approach, that is endeavouring to ascertain the personal conviction of a given judge in a given case, and an objective approach, that is determining whether he offered guarantees sufficient to exclude any legitimate doubt in this respect. 27 Later, both standards were further elaborated in Gregory v United Kingdom where the European Commission of Human Rights expressed its view that: If the possibility of bias on the part of the juror comes to the attention of the trial judge in the course of a trial, the trial judge should consider whether there is actual bias or not (a subjective test). If this has not been established, the trial judge or appeal court must then consider whether there is ' a real danger of bias affecting the mind of the relevant juror or jurors' (objective test). 28 Equally, in Remli v France, the European Court of Human Rights referred in this connection to the principles as developed in its case law regarding the independence and impartiality of tribunals. 29 The Court held in this case that, 'when it is being decided whether in a given case there is a legitimate reason to fear that a particular judge lacks impartiality, the standpoint of the accused is important but not decisive. What is decisive is whether this fear can be held to be objectively justified'. 30 The objective test of identifying impartiality of judges and jurors was also recognised by the Inter-American Commission on Human Rights which stated in Ramos v United States that the international standard on this issue requires the determination of 'whether there is a real danger of bias affecting the mind of the relevant juror or jurors. Where this bias may relate to a prohibited ground of discrimination, such as race, language, religion, or national or social origin, it may also implicate a violation of the principle of equality and non-discrimination.' 31

E. The Question of 'Intent' and 'Motive'
Closely related to stereotypical considerations and prejudiced attitudes of capital sentencing decision makers is the question why the alleged discriminator acted in a discriminatory way. In other words, the crucial issue of concern which is considered in this section is whether, as suggested by the standard of proof worked out by the US Supreme Court in the McCleskey death penalty case, the complainant (person sentenced to death) needs to prove a discriminatory motive or, more precisely, intent on the part of the respondent (alleged discriminator) in order to be successful in establishing direct discrimination in a particular capital sentencing process. 32 In some cases, a capital prosecutor, jury member or trial judge could argue, for instance, that there was no malice towards the complainant as an individual, but that their conduct was only the result of their wider generalisations about people with similar attributes (for example, 24  criminal stereotypes in the case of Afro-American defendants that they are inclined to commit grave crimes and that there is no real possibility for their correction). 33 It is necessary to clarify that the notions 'intent' and 'motive' in the context of discrimination law have different meanings and therefore should not be confused or used interchangeably as synonyms. 34 The intent can best be described as the goal of the discriminator's conduct and actions, 35 whereas the motive (which is the word that is usually conceived of in cases of direct discrimination) concerns the reason why the person perpetrated certain discriminatory action(s). 36 We can illustrate the role played by the motive in discrimination cases by the long-lasting trend in not applying the death penalty to women, which obviously favours female over male defendants. Such an inappropriate approach to the application of the death penalty appears to be present in almost all death penalty jurisdictions and reflects historical and widespread stereotypical attitudes towards vulnerability, moral inferiority and greater weakness of women (paternalistic approach). In spite of the particular severity and cruelty of this type of criminal sanction, it should be applied equally for men and women, regardless of sex and without relevance to the much greater number of men than women commit death-eligible crimes.
There is no legitimate reason -apart from a current state of pregnancy or maternity -that justifies the absolute exclusion of women from being sentenced to death. 37 In addition, this favourable treatment of female defendants in capital sentencing in comparison with male defendants is problematic also when it comes to proving direct discrimination in death penalty cases given that in most such cases the victim of discrimination (male capital defendant claiming discrimination on the grounds of gender) will have no direct access to evidence concerning the subjective motive of the discriminator; this makes it extremely difficult for the person sentenced to death to prove direct discrimination. What is of even greater concern is the possibility of influencing prosecutors', jurors' and judges' decisions in the capital sentencing process by various combinations and interactions of gender, race, age, socio-economic, employment and family status that perpetrators and/or victims of capital offences possess. 38 Accordingly, the discriminator's motive, which in fact reflects the subjective state of mind of the perpetrator, should not serve as a possible justification for direct discrimination in capital cases.
In some European states (including the UK and the Netherlands), it is now a general and established principle of law that motive cannot be relevant in proving direct discrimination. It follows from several domestic courts' decisions that unequal treatment cannot be justified by a discriminator's motive, intention, reason or purpose. 39 In light of these recent legal developments found in European countries, the US Supreme Court's reasoning in the McCleskey case, which is based on the Equal Protection Clause of the Fourteenth Amendment to the US Constitution, which places the discriminatory intent (as the state of mind) of jury members and judges at the centre of the concept of direct discrimination, appears to be faulty and outdated. These constitutional provisions impose on the capital defendant alleging a violation of equal treatment an excessive burden of 33 It is beyond debate that such arguments convey strongly stereotyped and racist notions. And the fact is that the capital defendant of Afro-American origin has been treated less favourably by the capital prosecution, jury and/or judiciary due to ascribed group characteristics which are, more often than not, inaccurate. 34 In the same vein, Gross distinguishes discriminatory 'intent' from motive or design and maintains that in most cases, such an intent 'must be inferred from circumstantial evidence, generally evidence that negates other possible explanations for a pattern or actions'. He also believes that 'judges can, if they wish, use an intent requirement to make a proof of discrimination difficult, or impossible'. Samuel R Gross, 'Race and Death: The Judicial Evaluation of Evidence of Discrimination in Capital Sentencing ' (1985) 18 Davis L. Rev. 1284-5. 35 In American jurisprudence the usual standard for proof of racially discriminatory intent requires ' a clear pattern, unexplainable on grounds other than race'; ibid 1295. 36 Schiek, Waddington and Bell (n 8) 227. proving 'the existence of purposeful discrimination' 40 perpetrated by decision makers in their case and, 41 as a corollary to this, that such purposeful discrimination had a discriminatory effect on them. 42 Without some kind of admission on the part of the capital prosecutor, juror or judge it is well-nigh impossible for a capital defendant to obtain the evidence of intentional or deliberate and conscious bias. 43 Indeed, with the McCleskey judgment the US Supreme Court rendered the entire American criminal justice system immune from judicial and wider public scrutiny when it comes to the issue of racial disparities and discrimination. 44

F. Difference in Treatment
A distinctive feature of direct discrimination's legal definition as contained in the EU non-discrimination directives is the requirement of less favourable treatment of the complainant. Hence, the relevant question in the legal context is whether any difference in treatment, however minor, can be taken to satisfy this legal criterion or whether a certain magnitude must be reached in order to be able to speak about such a difference in legal terms. Both the relevant text of EU non-discrimination directives and provisions of other international instruments on the prohibition of discrimination are silent on this issue. The lack of such legal qualification thus allows for national lawyers and judges to have differing legal interpretations as to the sufficient level of differential treatment. 45 However, in each capital case where the difference in treatment is established, domestic courts should very carefully consider whether the complainant actually suffered any detriment or disadvantage as a result of unequal treatment (for example, the capital defendant's legal defence was poor because of their weak financial, economic or social status or because the legal counsel approached the capital case involving an Afro-American defendant less seriously and acted less professionally than in similar cases dealing with white capital defendants).
The following section examines and compares some relevant proposals of proving discrimination in capital sentencing cases elaborated on in American federal and states' laws and presented in jurisprudence and legal scholarship, by shedding light on their benefits and shortcomings.

III. Initial proposals and recent US legislation, jurisprudence and scholarship on discriminatory application of the death penalty that alters the dominant US approach -Kentucky, North Carolina and New Jersey
In the aftermath of the McCleskey decision, there have been some congressional attempts in the United States to address the problem of proving racial discrimination in the administration of the death penalty by statistical evidence at the federal level which, in fact, all failed. 46 Thus, the initial federal legislative proposals 40 Blume, Eisenberg and Johnson believe that the requirement to prove -by direct evidence -the discriminatory intent (which is often invidious) on the part of decision-makers in capital cases under constitutional equal protection principles is not so much problematic itself but it is the perfunctory way the US courts treat available evidence and facts from which invidious discriminatory purpose may be inferred which is to blame for the reality that, in capital cases, racial discrimination claims have always failed. aiming to remedy the issue, commonly known as the federal Racial Justice Act and Fairness in Death Sentencing Act, have never been adopted or enacted as formal law. Both legislative proposals envisaged measures designed to provide those sentenced to death with the same right to challenge their individual sentences on the basis of racial discrimination as is afforded to individuals alleging racial discrimination in civil cases under federal employment and housing laws. The federal Racial Justice Act would enable a person sentenced to death to challenge their sentence if its imposition 'furthers a racially discriminatory pattern' 47 of death sentencing in the capital offender's jurisdiction. The proposed federal Racial Justice Act is built on a risk-based model of proof allowing the use of ' ordinary methods of statistical proof' 48 similar to those applied in the cases of racial discrimination claims in the context of jury selection. It also explicitly provided that such statistical evidence showing a racially disproportionate pattern of death penalty application should suffice to establish a prima facie case of racial discrimination and that proving any discriminatory motive, intent, or purpose on the part of the alleged perpetrator (individual or institution) is unnecessary. 49 The proposed law would give a state or federal entity the opportunity to rebut such prima facie evidence of a racially discriminatory pattern by furnishing ' clear and convincing evidence that identifiable and pertinent non-racial factors persuasively explain the observable racial disparities comprising the pattern'. 50 This implies that in the absence of such a rebuttal by the state the convicted person would be entitled to relief from their death sentence if their case fell within a category of cases with racial disparities to their detriment. In addition, this legislative proposal would also impose an obligation on death penalty jurisdictions to collect and maintain pertinent and sufficient statistical data on all potential capital cases so as to enable capital defendants and responsible state or federal entities to take necessary actions in the cases of future discrimination litigation under the Act. 51 The second proposal of US federal legislation, the so-called Fairness in Death Sentencing Act, was, in essence, very similar to the federal Racial Justice Act. It would also be predicated on a risk-based model of proof and thus allow an individual condemned to death to use statistics in order to prove a prima facie case of discrimination. The major differences between both proposals are that the Fairness in Death Sentencing Act contained more specific provisions on establishing a prima facie case by requiring that large and statistically significant racial disparities are presented by relevant evidence, 52 and that it would not require the state to collect any relevant statistical data. However, it would provide potential litigants with the possibility to obtain and present valid and official statistics made by public officials. 53 If, by means of statistical data, an inference was established that race was a basis for the death sentence, the government could rebut such inference only by a preponderance of the evidence. In the absence of such a rebuttal by the respondent government, the capital defendant would be relieved from their death sentence. Additionally, explanations and justifications made by the government that 'it did not intend to discriminate' or that 'the cases in which death was imposed fit the statutory criteria for imposition of death penalty' would not suffice to discharge its burden of proof, unless the government could provide a clear demonstration. 54 It is noteworthy that the two legislative proposals encountered pretty strong opposition in both the US Congress and Senate. 55 To prevail under the proposed Fairness in Death Sentencing Act, the capital defendant claiming the existence of racial discrimination in their capital case would need to prove by a preponderance of evidence either that race was a motivating factor in their individual case or that the death penalty was being 47 The relevant text of the proposed federal Racial Justice Act of 1989 is available as Appendix A to the article Baldus administered in a racially discriminatory manner in the jurisdiction in question. In principle, the Act would provide for the possibility to prove a prima facie case by direct evidence, such as racist remarks or admissions by key decision-makers in capital cases. However, the possibility of establishing a prima facie case of racial discrimination with statistical evidence is by far more important since it is very unlikely that direct evidence will be available and, therefore, it will be necessary for the court to draw inferences from circumstantial evidence showing racially discriminatory effects in certain death penalty jurisdictions. In addition, the proposed Act would require statistically significant evidence in substantiation of the claimant's allegations of discrimination. 56 Such substantial statistical disparity must account for the statutorily required aggravating factors appearing in all of the relevant capital cases. Hence, small or minor statistical or numerical disparities would fail to meet these requirements and would not be sufficient to support an inference of discrimination. As a result, in jurisdictions with a small death row population, statistical evidence alone, without further relevant evidence (either qualitative or quantitative), supporting the inference of discrimination would not be enough to establish a prima facie case. 57 Under the proposed Fairness in Death Sentencing Act, the respondent state could refute the plaintiff's claim of discrimination by several defences. First, it could demonstrate that the magnitude of statistical disparities presented by the claimant's evidence does not satisfy the legislative requirement of 'statistical significance' 58 to substantiate a prima facie case. Second, if the complainant succeeded in establishing a prima facie case, the respondent state would have the possibility of proving, by a preponderance of the evidence, that identifiable and non-racial factors persuasively explain the racial disparities. In doing so, the respondent state could use various reliable and generally accepted statistical procedures, methods and techniques taking into account statutory and non-statutory aggravating and mitigating factors that could ultimately clarify that the disparities, as shown by the claimant, are, in fact, based on objective and relevant non-racial or race-neutral factors and are not the product of unlawful discrimination.
In this sense, the statistical methods of proving discrimination, if employed correctly and appropriately, provide for more valid and just judicial evaluation of discrimination claims. Some congressional opponents contended that such a legislative proposal, if adopted, would require states to compile and maintain massive statistical databases. As suggested by some leading experts in the field, 59 information on relatively few factual circumstances -aside from data on the presence or absence of the statutory aggravating and mitigating factors -relating to each particular capital case is required in order to make the statistical analysis needed to support a claim of discrimination.
The need for the respondent state to employ the third defence available under the proposed legislation would arise only where the state could not disprove substantial statistical evidence showing that in a given category of death penalty cases, racial factors were affecting the sentences inflicted. If this was the case, the respondent state would still be able to prevail by demonstrating either that the complainant's case, due to factual differences, should not fall within the relevant category of capital cases affected by the observed pattern of discrimination or that the race of the capital defendant/victim is not captured by the demonstrated pattern of discrimination. The final possibility of the respondent state to successfully rebut the claim of discrimination supported by statistical evidence would be to show that, in fact, the state-wide evidence and the evidence from the complainant's own judicial district or circuit did not reflect a similar pattern of discrimination. 60 All things considered, one can readily conclude that the model of proof as contemplated in the proposed Fairness in Death Sentencing Act would provide adequate safeguards for the states as potential respondents from unjustified or indefensible discrimination claims provided that their death penalty systems are administered in an even-handed fashion.
In contrast to the legislative steps taken at the federal level, the project was successfully concluded by two state legislatures -Kentucky's and North Carolina's General Assemblies, where Racial Justice Acts allowing the use of statistical evidence in proving racial discrimination in the capital sentencing context were 56 Baldus, Woodworth and Pulaski suggest that rigid application of a statistical significance requirement 'may be appropriately relaxed' in those capital sentencing cases where 'because of a small sample size a demonstrated substantial racial disparity fails to achieve statistical significance and there is other convincing (qualitative or quantitative) evidence to support the inference of discrimination'. ibid 391. 57 See more on this issue in Baldus, Woodworth and Pulaski (n 46) 392. 58 Baldus, Woodworth and Pulaski noted that 'the level of statistical significance of a disparity in the death sentencing context depends primarily on the size of the disparity and the number of cases (sample size) involved in the analysis'. ibid 409. 59 ibid 397. 60 See also Baldus, Woodworth and Pulaski (n 46) 398. passed in 1998 and 2009 respectively. Under Kentucky's statute, 61 a capital defendant's racial discrimination challenge is limited to prosecutorial charging decisions. Capital defendants cannot challenge prosecutorial strike decisions, capital jury's decisions, or trial judges' decisions by arguing that these decisions led to discrimination against the defendant. To establish a prima facie case of discrimination, the statute allows a capital defendant to use statistical evidence, other (qualitative) evidence, or both. 62 Pursuant to the statute's provisions, the capital defendant has the burden of proving by ' clear and convincing' evidence that race was the basis of the prosecutor's decision to seek the death penalty. 63 Specifically, the capital defendant must state with particularity how their evidence supports a claim that race was a significant factor in the prosecutor's charging decision in their case. In other words, the capital defendant needs to show how a general pattern of racial discrimination in prosecutorial decisions to seek the death sentence affects their particular case. This requirement however, is somewhat ambiguous and allows for, at least, two different interpretations.
First, this statute's provisions may be construed broadly so as to require the capital defendant to merely demonstrate that their case falls within the category of cases embraced within the general pattern of racial disparities, for example that the prosecutor who sought the death sentence in their case was a part of a prosecutor's office that followed the state-wide pattern of discriminatory charging decisions (a risk-based approach). Second, in a stricter sense, the 'particularity' requirement could imply that the capital defendant must present the evidence of discrimination that is specific to their case (a causation-based approach). Indeed, such a strict interpretation would be counterproductive because it would largely frustrate the capital defendant's possibility to establish a prima facie case of discrimination by statistical evidence alone. 64 Therefore, the broader interpretation would be more reasonable. 65 If the capital defendant successfully demonstrates that the death sentence was sought because of race, the respondent state may offer its evidence to rebut the capital defendant's claims or evidence. The prosecutor is prohibited from seeking the capital sentence in a given case if the trial court finds that their charging decision was based on race. 66 Although Kentucky's Racial Justice Act provides some significant protections to capital defendants from potential racially discriminatory charging decisions, its model of proving discrimination contains some notable flaws. These include: 1) the impossibility to challenge jurors' and trial judges' decisions; 2) the temporal limitation of the possibility to challenge prosecutorial charging decisions to the pre-trial phase; and 3) the requirement that capital defendants need to present ' clear and convincing' evidence in support of their discrimination claims which is an unreasonably high burden of proof. Nevertheless, some observers note that the statute has had some limited impact on the prosecutorial charging decisions in death-eligible cases. 67 Another example of a state's legislation specifically addressing the issue of individual discrimination challenges in capital cases is North Carolina's recently repealed Racial Justice Act. 68 Under this statute, death row inmates who wished to raise discrimination claims were not required to produce evidence of purposeful discrimination in order to prevail in their individual case. North Carolina's law was very inclusive in its scope since it allowed capital defendants to challenge, by using statistical evidence, both the decision to seek and the decision to impose the death sentence. Thus, unlike Kentucky's Racial Justice Act, its provisions on discrimination claims applied to the entire capital sentencing process, including the capital jury's and trial judge's sentencing decisions. Its specificity was also that it entitled capital defendants to challenge prosecutorial peremptory strike decisions, including by using statistical evidence, if they considered that 'race was a significant factor in decisions to exercise peremptory challenges during jury selection'. 69 61 Alexander (n 46) 11. 62 The statute allows the capital defendant to use statistical evidence showing only state-wide disparities which means that statistics revealing disparities at the county, judicial district, or prosecutorial district level cannot be used in proving their discrimination claim. For a brief discussion on possible advantages and disadvantages of these provisions, see, Alexander (n 46) 11-13. 63 The suspect categories under the statute are both the race of the defendant and the race of the victim of capital offence; Alexander (n 46) 12. 64 See also, Alexander (n 46) 14. 65 ibid 12-13. 66 ibid 43. 67 ibid 14-15. 68 The statute originally provided strong and far-reaching protections to capital defendants from racially discriminatory capital sentencing decisions. However, in 2012 the act was revised in order to substantially weaken or limit some of the most capital defendant-friendly provisions; ibid 16. 69 ibid 17-18.
Many statistical studies, including the most comprehensive, rigorous and sophisticated, have shown that in the United States, race is a significant factor in all crucial decision-making phases of capital proceedings, from the point of indictment and charging decisions to jury selection and final sentencing decisions. 70 In its recent decision in the case of Robinson, the North Carolina Superior Court relied on the Michigan State University's study on jury selection, stating that the study can serve as powerful statistical evidence of racial bias in the North Carolina State's, 71 the former Second Judicial Division's and Cumberland County's capital sentencing systems, most notably in prosecutorial jury selection decisions at the time when Marcus Robinson was sentenced to death. Robinson's death sentence was consequently commuted to life imprisonment without possibility of parole. 72 Central for the Superior Court's decision was a landmark law passed in 2009, North Carolina's Racial Justice Act, under which capital defendants were allowed to challenge their sentence by showing that racial bias was a significant factor in decisions to seek or impose the death penalty. 73 This act was based on a risk-based model of proof of the type contemplated in the proposed federal Racial Justice Act and Fairness in Death Sentencing Act and did not require the complainant to provide further evidence showing factual discrimination in their particular case; in order to transfer the burden of proof from the complainant to the respondent state it was sufficient to show statistical evidence of statewide racial discrimination. However, efforts have been made lately to modify this unique act so as to change its relevant provisions on proving discrimination in capital cases and to require capital defendants to show discrimination in the county or prosecutorial district at the time the death penalty was sought or imposed. 74 Following these attempts, the North Carolina State House passed a new bill in June 2012 and a year later, in June 2013, but North Carolina's 2009 Racial Justice Act, was repealed by the Governor. 75 The main reason for this action was that 'nearly every person on death row, regardless of race, has appealed their death sentence under the Racial Justice Act' and that the act as such ' created a judicial loophole to avoid the death penalty and not a path to justice'. 76 Consequently, with the adoption of amended legislation concerning death penalty procedures, the provisions on proving racial discrimination in capital cases have been substantially weakened, to the prejudice of death row inmates. Thus, the current legal text provides that the burden to prove that race was a significant factor in decisions to seek or impose a death sentence rests on the capital defendant and that statistical evidence can be used as evidence relevant to establish a finding of a racially biased decision in the capital sentencing process; 77 however, the statistics alone are insufficient to establish a presumption or prima facie case that race played a significant role in prosecutorial or sentencing decisions. 78 Therefore, the act also allows for the use of other evidence of a qualitative nature, such as the sworn testimony of attorneys, prosecutors, law enforcement officers, judicial officials, jurors, or others involved in the criminal justice system. In addition, the capital defendant needs to clearly state, in particular, how the evidence provided supports their claim of racial discrimination. 79  submitted by the capital defendant, the respondent state may present to the court its evidence, not excepting statistical information. 80 Under this burden-sharing scheme, the burden of production falls upon the respondent state as soon as the capital defendant succeeds in establishing a prima facie case that race was a significant factor. 81 The respondent state must then rebut the claimant's case and dispel the inference of discrimination. In doing so, its simple non-discriminatory explanation, such as the commissioner's protestation that racial considerations did not play any part in the decisions to exercise a peremptory strike or seek or impose the sentence of death, will not be enough. The respondent state can use statistical or other evidence in rebuttal of the capital defendant's claims and evidence of discrimination. The statute provides that the respondent state may also offer the evidence showing that the state had ' any program the purpose of which is to eliminate race as a factor in seeking or imposing a sentence of death'. 82 The capital defendant bears the ultimate burden of persuasion and in considering whether they have met this burden the trial court will examine and weigh ' all of the admissible evidence and the totality of the circumstances'. 83 The Racial Justice Act provides for a single remedy: if the respondent state fails to rebut the capital defendant's prima facie evidence, the capital sentence must be vacated and the sentence of life imprisonment without the possibility of parole must be imposed. 84 It is important to note that not all US states' courts and judges chose to espouse the McCleskey approach in considering and evaluating detailed evidence (including statistics) presented when pursuing individual and structural capital sentencing discrimination challenges. For example, the New Jersey Supreme Court first addressed the issue of a structural challenge to the constitutional fairness of its death penalty system in State v Marshall, seeking to answer the operative question of whether the race of the defendant/victim 'played a significant part in capital-sentencing decisions in New Jersey'. 85 In its decision the Court eloquently rejected some crucial findings and arguments from the McCleskey ruling. It was particularly critical of the US Supreme Court's holding that ' apparent disparities in sentencing are an inevitable part of our criminal justice system' since such a statement indicates that ' absent purposeful discrimination, the Supreme Court apparently will not invalidate a death sentence on the basis of racial disparity, no matter how profound'. 86 It also specifically confronted the McCleskey Court's concerns and anticipation that, if the case was decided in favour of McCleskey, an unmanageable number of equal-protection claims would follow throughout the criminal justice system. 87 Although the New Jersey Supreme Court focused in Marshall mainly on structural challenges to the capital punishment system as a whole to find eventual evidence of unconstitutional discrimination (which was not found), it did not foreclose the possibility of recognising an individual claim of racial discrimination in a particular capital sentencing case.
Another such interesting example of a critical reaction towards the McCleskey decision comes from the Florida Supreme Court, where former Chief Justice Barkett in her dissenting opinion in Foster v State rejected the McCleskey ruling as inconsistent with the Florida Constitution and came up with an alternative approach to examining and proving racial discrimination with respect to prosecutorial decisions to seek the capital sentence. 88 Like the New Jersey Supreme Court, she emphasized that: 'Discrimination, whether conscious or unconscious, cannot be permitted in Florida courts. As important as it is to ensure a jury selection process free from racial discrimination, it is infinitely more important to ensure that the State is not imposing the ultimate penalty of death in a racially discriminatory manner.' 89 Obviously, it is possible to draw a parallel between the Barkett's model of proof and Justice Blackmun's approach to these issues put forward in his dissenting opinion in the proposals build on risk-based models and do not require evidence that the specific decision challenged by a discrimination claim was racially motivated. Thus, in order to establish a prima facie case according to Barkett's proposal, 90 it is sufficient for the claimant to prove that racial considerations affected relevant prosecutorial decision-making practices in a given state attorney's office. 91 To successfully rebut such a presumption of discrimination, the respondent state would then need to demonstrate that 'the practices in question are not racially motivated'. 92 If the respondent state fails to meet this burden of proof, it would not be permitted to seek the capital sentence in that case. 93 According to Chief Justice Barkett's belief, the evidence as applied in the context of proving discriminatory prosecutorial decisions to seek the death sentence should be construed as a broad concept including qualitative and, more importantly, quantitative evidence, such as statistical analysis, 94 or any other relevant information that might be suggestive of unlawful discrimination, such as 'the resources devoted to the prosecution of cases involving white victims as contrasted to those involving minority victims, and the general conduct of a state attorney's office, including hiring practices and the use of racial epithets and jokes.' 95 Barkett's proposal exhibits several virtues and is, in many respects, very close to the standard of proof entrenched in EU non-discrimination law. 96 First, it contemplates a burden-sharing approach as applied in discrimination claims in many other contexts and does not place an unreasonable burden of proof on the capital defendant. The second strength of this proposal is that it allows the trial court to evaluate all aspects of existing bias and to review all kinds of relevant evidence, including statistical data and analyses. The proposal's third advantage lies in the fact that it is possible to expect that such a model of proof would probably urge prosecutors' offices and other relevant state agencies to gather and maintain data on all death-eligible cases and make these data publicly accessible to enable their use in potential discrimination litigation. These records could also be applied for self-monitoring purposes, that is to scrutinise and eventually diminish/eliminate the involvement of conscious and unconscious racial bias in prosecutorial decision-making practices. Next, because it concentrates on discrimination challenges to prosecutorial decisions in a pre-trial phase, the model can be regarded as rather parsimonious. It would avoid the capital trials' and multiple post-conviction proceedings' expenses and harmful consequences for capital defendants and decrease the courts' workload. In addition, it would not require the state to establish and maintain impossibly large and expensive databases. Lastly, given its apparent focus on discrimination challenges to prosecutorial decisions to seek the death sentence, it would require only prosecutors to explain and defend their challenged decisions, whereas the same would not be asked from capital jurors and trial judges, who are normally less likely to be in a position to do this.
The major drawback of the standard of proof proposed by Chief Justice Barkett is that it seems to ignore a judicial and, in particular, capital jury's decision-making phase and is, therefore, too narrow in its possible application. 97 There are many veracious statistical and narrative sources, including depositions of those serving on capital juries, 98 as well as experimental studies which confirm that racial bias often plays a substantial role in the capital jury's decisions. Even more significantly, it was found in Andrews v United States by virtue of clear and convincing evidence that unlawful racial bias entered the capital jury's decisionmaking process in this particular case. 99 Therefore, the application of the same standard of proof should necessarily be extended to discrimination claims targeting biased capital juries' decisions. 100 Some interesting proposals for identifying an appropriate model of proof to be applied in individual capital sentencing discrimination claims have been presented in legal scholarship as well, mainly in response to the existing legislative and judicial models. These alternative models elaborate on previous experiences 90 It is important to note that the Barkett's approach covers both the race of the capital defendant and the race of the victim; ibid 468. 91 ibid 467-68. 92  and developments and seek to remedy some major deficiencies and failures of the evidentiary approaches embraced in the proposed or existing legislation and in jurisprudential interpretations. Alexander's recent proposal, 101 in particular, deserves a more detailed review and analysis.
This proposal places on the capital defendant the burden to prove ' clearly and convincingly' their claim that 'race was a significant factor in decisions to seek or impose the sentence of death in the county, prosecutorial district, or state at the time the death sentence was sought or imposed'. 102 Here it must be pointed out that, in general, the standard of ' clear and convincing' evidence requires a clear-cut and direct proof of discrimination (for example, an admission on the part of an alleged discriminator). This type of proof is hardly ever available in discrimination cases, and to require the complainant to present this evidence thereby precludes the hearing court from drawing inferences of discrimination from available primary facts. 103 Therefore, this aspect of the proposed model's legal standard of proof is too onerous and somewhat unfair for the capital defendant. In addition, the author does not clarify or exemplify what proof is able to meet the ' clear and convincing' evidence requirement, although the proposed text does mention some examples which might be illustrative of what can be considered ' clear and convincing' for the purposes of the statute, such as statistical evidence (either alone or in combination with other available evidence), and sworn testimonies of attorneys, prosecutors, law enforcement officers, judicial officials, jurors, or others involved in the criminal justice system. 104 According to Alexander's model of proof, the defendant claiming racial discrimination in their capital case would need to provide powerful evidence in support of their claim in order to get the government before the court to conduct an evidentiary hearing. 105 The weakest and most unfounded claims of racial discrimination could be dismissed by the court without having to seek any response from the government, thereby decreasing unwarranted discrimination litigation costs. 106 In the case of the defendant's strong statistical evidence, however, the respondent government would have a legitimate possibility to rebut such evidence by showing that in the claimant's individual case statistical disparities in death sentencing can reasonably be explained by non-racial factors. 107 Accordingly, the claimant would carry the burden of establishing by a preponderance of the evidence a prima facie case of discrimination. If the claimant succeeded in establishing such a presumption of discrimination, an inference of discrimination would arise and trigger a shift in the burden of proof. Then the respondent government would be required to articulate a legitimate and non-discriminatory reason to rebut the established inference of discrimination. Once the government successfully showed a non-discriminatory reason for its decision, the burden of proof would shift back to the claimant to clearly demonstrate that the respondent government's articulated non-discriminatory reason for its action was a mere excuse for unlawful discrimination. 108 Risk-based models of proof, as described above in this section, clearly provide for a causal inference. This means that once an unrefuted prima facie case has been established by the claimant, the provided evidence sustains an inference that race, ethnicity, gender, sexual orientation and/or socio-economic status of the capital defendant and/or victim of a capital offence substantially influenced decisions on seeking or imposing the death sentence in some death-eligible cases implicated by the statistical analysis. The unrebutted statistical evidence of discrimination thus creates a presumption that prohibited ground(s) of discrimination adversely affected each case so implicated. This presumption is irrefutable under a risk-based model of proof. As a consequence, all cases implicated by such statistical evidence are potentially eligible for relief and, therefore, definitive proof of a causal connection between the prohibited discrimination ground(s) and the adverse decision in any specific case within the relevant group of cases is unnecessary. 109 101 Alexander (n 46) 23-41, 46-47. 102 Alexander (n 46) 27. 103 See more on this point in section IV B below when discussing the applicability of the concept of shared burden of proof in the capital sentencing context. 104 Alexander (n 46) 46. 105 The suspect categories captured by the proposed model law are: 1) the race of members who are removed from capital juries through prosecutorial peremptory challenges; 2) the race of capital defendants; and 3) the race of victims of capital crimes; Alexander (n 46) 34-39. 106 The proposed statute would allow capital defendants to raise their claims of racial discrimination in both the pre-trial phase and post-conviction proceedings; Alexander (n 46) 46. 107 The proposed legislation would allow capital defendants to ground their discrimination claims on statistical evidence showing capital sentencing disparities in a particular county, state judicial appellate district, or state; Alexander (n 46) 30. 108 Alexander (n 46) 27. 109 Baldus, Woodworth and Pulaski (n 46) 398.
Causation-based models of proof, by contrast, attempt to identify those cases in which decisions to seek or impose the death penalty were actually adversely affected because of biased considerations. Causationbased models thus require an additional judicial examination of the specific facts present in the claimant's case. Under the causation-based model the respondent state is provided with an additional and final defence against unrebutted statistically-based evidence; that is, it can overcome the presumption of discrimination as it applies to the claimant's case if it is able to prove by objective evidence concerning similar cases that race, ethnic origin, gender, sexual orientation, socio-economic status and so on: 1) were not a but-for factor; or 2) that such characteristics did not play a substantial or significant role; or 3) that they played no role at all in the claimant's case. Indeed, the latter two approaches would impose a much more demanding burden of proof on the respondent state than the but-for model which is the most conservative model of proof. 110 Moreover, the evidentiary approach based on the causation model of proof usually requires the respondent to shoulder a higher burden of proof, such as a requirement to disprove the authoritative unrefuted presumption of discrimination by clear and convincing evidence. 111 It is worth noting that while the causation-based models of proof seek, on the one hand, to compensate for one of the greatest deficiencies of the risk-based models, for instance, allowing some offenders to evade their execution even though personal characteristics in fact played no, or perhaps merely a minor, role in their specific capital cases, they are, on the other hand, simply incapable of doing away with a peril that some capital sentences are affected in part by discriminatory factors and that these considerations will not be identified and, subsequently, some persons wrongfully sentenced to death will be executed. 112 Accordingly, if one wants to diminish unlawful discrimination in certain capital sentencing systems or jurisdictions as much as possible and, in particular, to prevent executions arising from discriminatory capital sentences, risk-based models of proof should be by far the preferred legal alternative, although such methods of proof may eventually be criticised for providing too much protection to capital defendants and imposing too heavy an evidentiary burden upon the state.

IV. What Standard of Proof in the cases of discrimination in capital cases?
Perhaps the most challenging task for the capital defendant as a victim of discrimination is to provide sufficient evidence to prove that the death penalty was sought or inflicted on them because of their race, national or ethnic origin, gender, religion or belief, socio-economic status, sexual orientation and so on. For judges deciding such discrimination cases, it may be equally difficult to define the level of impact or extent to which the less favourable treatment was caused or influenced by one of the prohibited grounds of discrimination or their combination/intersection. Such unlawful discrimination may have played a certain, though not exclusive, role in giving rise to unequal treatment. Over time, as fewer and fewer cases involved overt discrimination, legal practitioners faced the challenge of unearthing and proving covert, hidden and subtle forms of discrimination, also in the context of the death penalty application. In light of such real-life developments it soon became self-evident that clever tactics would be necessary in order to 'smoke out' shrouded discriminatory attempts and provide the adequate evidence of direct discrimination.
One possible measure for collecting important evidence which could be applicable to the cases of discriminatory application of capital punishment could include, for example, a questionnaire prescribed by domestic legislation in which the complainant sets out the incidents, circumstances and situations in which they believe they experienced unlawful discrimination. The respondent is then requested to provide their explanation in response to complaints written in the questionnaire. If such an informal procedure is later followed by a tribunal or other formal proceedings, the litigants then have the possibility of invoking their statements and responses from the questionnaire. What is more, the respondent's failure to provide their response may result in negative inferences drawn by the court. But, naturally, prosecutors, jury members and judges acting in capital cases will be astute enough to evade any incriminating responses.
In many discrimination cases related to capital sentencing processes, national courts are faced with some evidence suggesting that discrimination may have taken place, but which is insufficient to definitively prove that unequal treatment was the real reason for the alleged discriminator's actions. Indeed, the approach 110 See ibid 399. 111 See ibid 400-401. 112 Baldus, Woodworth and Pulaski state that the risk of executions following some discriminatory death sentences exists even if one applies the most stringent test under the causation-based models of proof, ie the requirement for the respondent state to provide a convincing and clear proof that race, gender, sexual orientation, etc played no part whatsoever in the capital defendant's case; ibid 401-402.
of domestic courts as to whether to accept certain evidence as sufficient to establish a prima facie case of discrimination will depend on the legal area in which the complaint is grounded. Thus, the standard of proof in criminal law will typically be higher than in civil law. Likewise, the rule on the reversal or sharing the burden of proof as a measure to alleviate the problems of proof in discrimination litigation will normally be applicable in the civil law context but not in criminal law cases involving a discriminatory motive. 113 In capital cases in particular, it is rare that the court is provided with evidence unequivocally showing discriminatory reasons for the respondent's decision, conduct or act. In most such cases, judges will need to draw inferences of discrimination taking into account the totality of the available evidence. 114 This will enable the court to arrive at the conclusion regardless of whether such an interpretation can be accepted as the most probable explanation. To put it another way, in the absence of express evidence of discrimination in capital cases, courts should draw inferences based on the balance (that is, preponderance) of the evidence. 115 This section first reflects on the international legal status of such a standard. Then it discusses to what degree the standard of proof contained in modern European non-discrimination law is appropriate or applicable to the context of capital proceedings.

A. International Legal Character of the Standard of Proof
For the purposes of further discussion, it is first of utmost importance to determine the international legal status of the standard of proof defended and detailed in this article. The standard according to which the evidentiary issues should be judged and assessed lies at the heart of discrimination law. To be in a position to clearly define the international legal nature of such a standard and, consequently, international legal obligations of death penalty jurisdictions as far as unlawful discrimination in death penalty cases is concerned, including the application of adequate evidentiary rules and procedures which must necessarily provide for procedural equality and fairness, one needs to carefully consider primary sources of international law as enumerated in Article 38 of the Statute of the International Court of Justice, relevant international customary and human rights treaties' provisions, and general principles of law recognised by civilised nations. States, as members of the international community, are obliged to abide not only by provisions of international conventions to which they are parties and by norms of customary international law but also by general and fundamental legal principles that are common to all major legal orders.
While it is hard to contend that the standard of proof as elaborated in this article fulfils the criteria which are required to identify and ascertain the norms of customary international law, it is more certain that such a standard enjoys -at least to a certain extent -the status of a general principle of international law within the meaning of Article 38(1)(c) of the Statute of the International Court of Justice. General principles of law recognised by civilised nations include some principles which are characteristic of every judicial proceeding and adjudication. 116 For instance, the principle of independence and impartiality of an adjudicative body and its unbiased and non-discriminatory performance of the adjudicative function can be placed within this category, as can be the principle of equality and hearing of both parties participating in the proceeding. The same is true of the procedural principle according to which the judgment must always be reasoned and rest on good grounds.
Moreover, the standard of proof proposed herein can be said to be implicitly comprised in the relevant provisions of all core international human rights instruments adopted at both the universal and regional levels, such as the International Covenant on Civil and Political Rights, the International Convention on the Elimination of All Forms of Racial Discrimination, the Convention on the Elimination of All Forms of Discrimination Against Women and the European Social Charter. The competent human rights bodies established by these international treaties have interpreted this standard of proof in much the same way, that is, pointing towards the modern standard of proof featured in the advanced EU non-discrimination law (including the relevant case law of the Court of Justice of the European Union and EU directives prohibiting 113 For example, the rule on sharing the burden of proof will not normally apply in criminal proceedings where the perpetrator is prosecuted for a certain crime which was motivated by race -or sexual orientation-related stereotypes and prejudice (hate crimes). This is partly because a higher standard of proof is needed to establish criminal liability, and partly because it would be difficult to require a perpetrator to prove that they did not hold a discriminatory motive, which is entirely subjective in its nature. discrimination) as well as in the domestic anti-discrimination legislation of all 28 EU Member States and beyond. Hence, it is reasonable to conclude that the standard of proof, as outlined in this article, is indeed part of international law.

B. The Shared Burden of Proof in Contemporary Discrimination Law and the Applicability of the Concept in the Capital Sentencing Context
Normally, it is for the claimant to convince the court or other adjudicative body that there has been a serious violation of human rights. Unlike other cases of human rights violations however, discrimination cases require a slightly different approach when it comes to evidentiary issues. The reason for this is, as discussed above, because the motive behind such unequal treatment is and remains hidden in the head of the discriminator and therefore, it may be particularly difficult for the claimant to prove that the less favourable treatment they suffered was really on the basis of a particular ground of discrimination protected by law. This is why in most cases the claim of discrimination consists mainly of ' objective inferences related to the rule or practice in question'. 117 However, the judges deciding on the issue of the discriminatory application of the death penalty must be convinced that the only reasonable explanation for seeking or imposing the death penalty lies in the defendant's and/or victim's protected characteristic(s). The alleged discriminator will usually be in possession of information needed to prove or disprove a claim of discrimination before the tribunal, including in capital sentencing processes. Therefore, the rules on providing evidence in discrimination cases should be adapted to this specific situation in the form of sharing the burden of proof with the respondent (alleged perpetrator). The purpose of the rules on alleviation of the burden of proof is to enhance the effectiveness of statutory provisions on protection against discrimination. The concept enables courts to deal with discrimination phenomena in the light of the discriminatory effects generated by a certain rule, act or practice.
Having in mind the aforementioned concerns and in order to somewhat ease the claimants' troubles with providing sufficient evidence of discrimination, national legislation on non-discrimination should allow the procedural possibility of proving discrimination through appropriate rules, evidence criteria and burdens of proof, derived from the understanding that the victims of discrimination are usually at a disadvantage and would not be able to defend their rights in the courts unless special care is taken as to their procedural rights. As a result of this understanding, some national legal systems introduced the rule on shifting the burden of proof from the claimant to the respondent once a prima facie case of discrimination has been established. This specific principle in the evidentiary procedure is a characteristic of EU non-discrimination law as well. 118 Indeed, one of the key issues in shifting the burden of proof is to determine how much evidence needs to be presented by the complainant to trigger a shift of the burden to the respondent. 119 Provisions on the burden of proof constitute one of the core elements of non-discrimination legislation. It is therefore very important for a legislator to strike the proper balance between the petitioner's and respondent's evidentiary onuses in order to provide for an effective non-discrimination law. In other words, non-discrimination legislation must not place too low a burden on the respondent or too high a burden on the claimant and vice versa. Only a meritorious claim of unlawful discrimination should form a legitimate basis for a petitioner's success and the respondent should be given a fair chance to rebut the petitioner's evidence.
The rule on the distribution of the burden of proof between the claimant and respondent in discrimination cases is part and parcel of the EU legal framework as the relevant provisions can now be found in all core EU non-discrimination directives. According to these provisions: Member States shall take such measures as are necessary, in accordance with their national judicial systems, to ensure that, when persons who consider themselves wronged because the principle of equal treatment has not been applied to them establish, before a court or other competent authority, facts from which it may be presumed that there has been direct or indirect discrimination, it shall be for the respondent to prove that there has been no breach of the principle of equal treatment. 120 117 Handbook on European non-discrimination law (n 113) 124. 118  In the context of the death penalty's application, such a definition would signify that where a capital defendant substantiates by prima facie evidence facts from which it may be presumed that they have been treated less favourably on any of the prohibited discriminatory grounds, it shall be for the respondent to prove that this treatment is justified by objective considerations other than the complainant's protected status. Accordingly, the claimant would need to carry the initial burden of presenting factual evidence that supports the assumption that direct discrimination has occurred. As soon as such factual evidence is produced, the respondent would have to prove that the controversial decision of a prosecutor, jury or judge was justified on the basis of objective reasons unrelated to the prohibited grounds of discrimination. 121 Similarly, in their rather comprehensive and critical review of post-McCleskey racial discrimination claims focusing on prosecutorial decisions to seek the death penalty, Professors Blume, Eisenberg and Johnson arrived at the following conclusion: [I]f the defendant establishes a prima facie case of discrimination, by which we mean a showing by statistical or other evidence that discrimination is more likely than not to have occurred, then the burden of production shifts to the prosecutor, who must then articulate some legitimate, non-discriminatory reason for seeking the death penalty. If the prosecutor fails to assert such a reason, the defendant prevails. If the prosecutor produces evidence creating a genuine issue of fact, then the defendant, retaking the burden of persuasion, must establish discrimination by a preponderance of the evidence. 122 Indeed, the point must be made that EU non-discrimination directives provide only basic minimal standards to be implemented and applied by the EU member states. Consequently, national legislation may introduce rules of evidence which are more favourable to claimants. 123 The EU non-discrimination law thus allows the burden of proof to be shared between the litigants. According to this approach, the claimant needs to submit sufficient evidence suggesting that differential treatment may have taken place. Such evidence will raise the presumption of unlawful discrimination, which the respondent then has to successfully refute in order to relieve himself of this burden. Consequently, once the complainant is able to present facts from which it can be presumed that discrimination may have occurred, the burden of proof shifts to the respondent (alleged perpetrator) to prove otherwise. The phrase 'may be presumed' in the above concept of shifting the burden of proof is of paramount importance in the process of establishing discrimination. These words imply that it is for the claimant who complains of race, sex, religious and sexual orientation discrimination to prove on the balance of probabilities facts from which the court or tribunal may conclude, in the absence of an adequate explanation, that the respondent has committed an act of discrimination against the claimant which is unlawful by virtue of relevant legal provisions. If the claimant fails to produce the evidence proving such facts, their complaint will be unsuccessful. Obviously, it will be almost impossible to find and submit direct evidence of unlawful discrimination in capital cases. 124 The fact is that very few, if any, prosecutors, jury members or judges would be prepared to admit such discrimination, even to themselves. In this respect, the guidance 121 Interestingly enough, in Foster v State, Chief Justice Barkett proposed rather analogous standard for proving racial discrimination in the death penalty decision-making context: 'A party asserting racial discrimination in the State's decision to seek the death penalty should make a timely objection and demonstrate on the record that the discrimination exists and that there is a strong likelihood it has influenced the State to seek the death penalty. Such discrimination conceivably could be based on the race of the victim or on the race of the defendant. Once the trial court determines that the initial burden has been met by the defendant, the burden then shifts to the State to show that the practices in question are not racially motivated. contained in the UK Court of Appeal's judgment delivered in IGEN Ltd & Ors v Wong, saying that 'in some cases the discrimination will not be an intention but merely based on the assumption', can be of particular relevance. 125 Thus, in order to come to a proper outcome, a court or tribunal will usually need to analyse such primary facts and draw inferences therefrom. The above wording 'may be presumed' indicates that at this stage of the evidentiary procedure, the court or tribunal does not have to come up with the ' definitive determination that such facts would lead it to the conclusion that there was an act of unlawful discrimination', but rather that it should look at the primary facts presented to it to see 'what inferences of secondary fact could be drawn from them'. 126 This requires that the court or tribunal, when it considers what inferences or conclusions can be drawn from the primary facts, assumes that 'there is no adequate explanation for those facts'. 127 Hence, the court or tribunal should appropriately examine ' any inferences that it is just and equitable to draw' in a particular case. 128 Such inferences may include, for instance, the respondent's evasive or equivocal responses to a questionnaire (as described earlier in this section) or their failure to comply with any existing and relevant code of practice. 129 In this context, Bangalore Principles of Judicial Conduct adopted in 2002 which lay down fundamental, though non-obligatory, standards on the proper administration of justice, can surely be of great help and relevance. These Principles provide that ensuring impartiality of the courts and equality of treatment to all before the courts is essential to the proper discharge of the judicial office. This applies 'not only to the decision itself but also to the process by which the decision is made' and therefore ' a judge shall perform his or her judicial duties without favour, bias or prejudice'. 130 Where the claimant has been successful in proving the facts from which conclusions could be drawn that any ground of unlawful discrimination gave rise to less favourable or disadvantageous decisions made by the capital prosecutor, jury members or trial judge and which, subsequently, adversely influenced the outcome of the capital sentencing process in question, the burden of proof moves to the respondent. It is then for the respondent to prove that they did not commit the act of unlawful discrimination by adopting the controversial decision. To successfully discharge this onus the respondent must prove, on the balance of probabilities, 131 that the decision impugned by the capital defendant was not made on the grounds of a discriminatory factor. This means that the court or tribunal will have to assess not only 'whether the respondent has proved an explanation for the facts from which such inferences can be drawn', 132 but also whether 'it is adequate to discharge the burden of proof on the balance of probabilities' 133 that the decision at issue was not by any means affected by the capital defendant's and/or capital crime victim's race, national or ethnic origin, or other legal protected characteristic. Normally, the facts and information necessary to prove an explanation will be in the hands of the respondent. Therefore, it is reasonable for the court to require cogent evidence to discharge the respondent's onus of proof. 134 In the cases where a claimant argues and succeeds in furnishing sufficient evidence of facts suggesting the needed probability of discrimination in the capital sentencing decision-making process, the major inquiry to be made by the court will be to establish if the capital defendant's and/or victim's personal characteristics played any role whatsoever in the capital prosecutor's, jury members' and/or trial judge's decision-making 125 130 All these principles shall be strictly applied, particularly in death sentencing processes; Bangalore Principles of Judicial Conduct, E/ CN.4/2003/65, annex 3, 4, 6, 7. 131 By way of contrast, Lyon suggests that the respondent (ie prosecution) must demonstrate, 'beyond reasonable doubt', that an 'impermissible appeal to bias did not affect the fairness of the defendant's trial'. Lyon (n 15) 320. Indeed, one should not forget that discrimination litigation, even in the criminal sentencing context, are civil cases in their nature and therefore requiring the litigants to meet such a high evidentiary threshold as is standard 'beyond reasonable doubt' and which is normally pursued in criminal trials would be at least unreasonable, if not illegitimate. Against that backdrop, the standard of proof which is based on the preponderance of the evidence seems to be a far more appropriate and just solution. In the same vein, the Committee on the Elimination of Discrimination Against Women pointed out in V. K. v Bulgaria that in civil proceedings relating to domestic violence complaints the requirement that the act of domestic violence must be proven 'beyond reasonable doubt' constitutes an ' excessively high' standard of proof which is 'not in line with the Convention on the Elimination of All Forms of Discrimination against Women, nor with current anti-discrimination standards which ease the burden of proof of the victim'; V. K. v Bulgaria (20/2008), CEDAW/C/49/D/20/2008 (2011) para 9.9. 132 IGEN Ltd & Ors v Wong (n 123) para 76(12). 133 ibid. 134 For example, the tribunal will have to carefully consider the respondent's explanations for not adequately dealing with 'the questionnaire procedure and/or code of practice' if these should have been applied in a given case; ibid para 76(13).
activities and, in particular, in the outcome of the criminal process. Indeed, serious, exact and consistent factual elements (for example, statistical data) should provide for a clear establishment of the presumption of the existence of direct discrimination. But somewhat less compelling features like a lack of adequate transparency in terms of the possibility to identify non-discriminatory capital sentencing decision-making or the respondent's behaviour or conduct that seems objectively unreasonable should also shift the burden of proof to the respondent. As a result of such an approach, a finding of discrimination can be made even where this was not the exclusive reason for the capital prosecutor's, jury members' and/or trial judge's decision or motion. A direct distinction is also present where the grounds of discrimination were only part of the reason for the impugned decision or action. 135 Consequently, if it is not proven, in a particular case, that the discrimination does not exist and/or that the respondent did not act in contravention with the legal prohibition of discrimination, then they must be held responsible for direct discrimination based on the relevant ground(s) of discrimination. Put another way, the respondent can disprove the presumption of discrimination if they succeed in proving either that the claimant is not actually in the same, similar or comparable situation to their comparator or that disparate treatment is not based on the protected characteristic but other objective differences. 136 The following example may serve as an illustration of how, pursuant to EU non-discrimination law, the burden of proof should be distributed between the litigants in cases dealing with discrimination in the application of the death penalty. First, the claimant alleging that they were sentenced to death because of their and/or the capital crime victim's race, national or ethnic origin, gender or such other characteristics will need to prove that they were subjected to capital punishment, whereas capital defendants of another race, ethnicity and so on were not.
Secondly, they will need to show that their relevant comparator is in a similar or comparable situation as regards the nature of a capital crime committed, their criminal record and the degree of their culpability, amongst other factors. If the claimant is able to adequately present such circumstances and characteristics this suffices for the presumption being raised that the difference in the final conviction and punishment could only be explained by discrimination against the claimant's personal characteristics. At this point the onus of proof falls on the respondent's shoulders and it is for them to rebut the presumption of discrimination, for example by showing that white and non-white capital defendants were not actually in a comparable situation because the circumstances and characteristics of both cases are substantially different in their nature. Moreover, the respondent can discharge themselves of the burden of proof also by demonstrating that the difference in the outcome of both capital cases can be explained by objective factors, unrelated to a personal characteristic of the capital defendant and/or the victim of a capital crime.
As for the core international and regional legal instruments explicitly prohibiting discrimination on various grounds, such as the Koutropoulos v Greece 138 and Timishev v Russia. 139 In the cases before it, the Court will generally look at the available evidence as a whole, bearing in mind that in most similar cases much of the information which is really relevant to prove the claim of discrimination will be in the hands of the respondent government. If, in its view, the facts presented by the applicant are sufficiently credible, veracious, reliable and consistent with the available evidence, the Court will normally accept them as proven, unless the respondent government provides a cogent alternative explanation. 140 In its case law the Court has also made express reference to the rule on shifting the burden of proof as provided in the EU non-discrimination directives. Yet, it has approached the possibility of applying the same principle within the context of the Convention in a more prudent and conservative manner. It seems that the Court has been unwilling to espouse and stick to the firm rule that the burden of proof shifts from the applicant to the respondent once a prima facie case of discrimination is established by the applicant, but rather has chosen more discretionary steps allowing for inferences of discrimination to be drawn from the specific facts and the particular part of the Convention to be applied in each individual case. 141 Considerations of the Inter-American Commission on Human Rights are even more interesting since they concern -albeit more indirectly than directly -evidentiary issues exactly in the cases dealing with discrimination in the death sentencing context. In Celestine v United States, both the US courts and the Inter-American Commission on Human Rights emphasized that proving a certain pattern of (racial) bias and discrimination as such is insufficient to establish a prima facie case of discrimination in a particular case such as to shift the burden of proof to the respondent government. 142 The cases such as Andrews v United States, 143 Roach and Pinkerton v United States, 144 Ramos v United States and Landrigan v United States, 145 however, demonstrate examples of situations where the Inter-American Commission was willing to draw an inference of discrimination.
In Andrews v United States, the Inter-American Commission accepted the petitioner's evidence of racial bias as sufficient to substantiate a prima facie case of racial discrimination in the capital sentencing process and trigger a shift in the burden of proof. 146 Since the US government was unable to discharge this burden, the Inter-American Commission found that the United States violated Article II of the American Declaration of the Rights and Duties of Man which guarantees the right of everyone to equal treatment at law without discrimination as to various grounds, including race. 147 For the Inter-American Commission the napkin (note) with flagrant racist content (there were words written in black stating 'hang the nigger's' and a figure drawn in black hanging therefrom) 148 found among the jurors as well as other factors indicating racist views of jury members (for example, the jury in this case was all-white and it contained several Mormons whose religious faith was based on the racist ideology and tenets believing that black people are inferior beings), and the fact that all this was known to the trial judge, who did not question the jury members about the incident but only admonished the jurors to ignore it and proceeded to trial with the same members of the jury, reflected ' ample evidence' of 'racial bias'. 149 In addition, the United States did not dispute these facts before the Inter-American Commission.
The may be executed resulted in ' a pattern of legislative arbitrariness throughout the United States'. 150 As a result of such a legislative scheme, the severity of punishment was not based on the nature of the crime committed but on the location where it was committed. The Inter-American Commission pointed out that states should not be granted any discretion when it comes to the possibility of the imposition of the death penalty upon juveniles and their execution. 151 Such a legal situation is, in the view of the Inter-American Commission, contrary to Article II of the American Declaration of the Rights and Duties of Man, which provides for the right to equality before the law.
In another case, Roberto Moreno Ramos v United States, the Inter-American Commission took the view that consideration of the nationality of the accused, when irrelevant, in sentencing death penalty cases amounts to a violation of the right to equality before the law. 152 The petitioner, who was a national of Mexico, claimed that criminal proceedings against him were unfair due to serious trial deficiencies, including the fact that during the punishment phase of his trial the prosecution made discriminatory comments related to his nationality, thereby depriving him of the right to equality before the law. 153 Based on the facts of the case and the evidence provided, the Inter-American Commission concluded that 'viewed objectively and in the context of the circumstances of Mr. Moreno Ramos' crime and the purpose of the sentencing hearing more broadly, there is a real danger that Mr. Moreno Ramos' nationality was considered by the jurors in determining his punishment'. 154 The Inter-American Commission was of the opinion that numerous aspects of the petitioner's punishment hearing, including the discriminatory manner in which the prosecutor referred to the petitioner's status as a foreign national, suggested such a conclusion. According to the Inter-American Commission, in the present case the nationality of the accused is a factor which was completely 'irrelevant to and unconnected with the issues under consideration at the punishment phase of his trial, raising a particular danger that this evidence would be relied upon as a consideration in determining an appropriate sentence'. 155 Since no steps were taken, by the trial judge or otherwise, to clarify that the jurors were not to consider the petitioner's nationality as an element in deciding upon his punishment, there existed ' a real possibility' that the petitioner's nationality had been taken into consideration by the jury members in determining whether he should be executed for his crime. 156 The Inter-American Commission thus decided that the domestic criminal justice system failed to afford the petitioner 'his right to be tried by an impartial tribunal as well as his right to equal protection of the law without discrimination'. 157 More recently, in 2011, the Inter-American Commission considered the question of whether a procedural distinction in a capital sentencing case which resulted in the petitioner not having his sentence reviewed constituted discrimination in violation of Article II of the American Declaration of the Rights and Duties of Man. 158 This case is particularly interesting since the alleged differential treatment was based on grounds other than race, nationality, sex, language or creed. The American Declaration contains in Article II a nondiscrimination clause which comprises an open-ended list of grounds of discrimination. Thus, the discriminatory treatment related to the possibility of obtaining judicial review can be placed within the formulation of ' any other factor'. The petitioner was 'part of the group of persons that was denied access to the benefit of review of a sentence that resulted from an unconstitutional procedure, on the sole basis of the procedural stage of his case'. 159 The Inter-American Commission rejected the US government's arguments for differential treatment, namely those of judicial economy, certainty, and legal security and underlined that such justifications need to be 'balanced against the nature of the individual rights at issue -which may involve the protection of life, liberty and physical integrity'. 160 Finally, the Inter-American Commission concluded that the distinction applied to the petitioner's case was not reasonable and legitimate, and that the differential 150  legal treatment received from the courts amounted to inadmissible discrimination. The Inter-American Commission therefore found the US government responsible for violating the petitioner's right to equal treatment before the law by denying him, in an unjustified and discriminatory fashion, the determination of his basic rights including, possibly, the right to life itself. 161 In addition to regional human rights courts and commissions, other international human rights treaty bodies have also developed some relevant guidelines concerning the burden of proof in discrimination cases and its distribution between litigants, including in the criminal justice administration context. For instance, in relation to discrimination against non-citizens, the Committee on the Elimination of Racial Discrimination recommended that the States Parties to the International Convention on the Elimination of All Forms of Racial Discrimination: Regulate the burden of proof in civil proceedings involving discrimination based on race, colour, descent, and national or ethnic origin so that once a non-citizen has established a prima facie case that he or she has been a victim of such discrimination, it shall be for the respondent to provide evidence of an objective and reasonable justification for the differential treatment. 162 The same approach encouraging a more balanced sharing of the burden of proof between the claimant and the defendant in civil proceedings involving racial discrimination was taken by the Committee in its concluding observations addressed to particular States Parties. 163 The issue of a proper distribution of the burden of proof in proving discrimination has also been addressed by the European Committee of Social Rights, an expert body established to monitor the implementation of the European Social Charter (revised) in States Parties. For instance, in its decision on a collective complaint in Mental Disability Advocacy Center (MDAC) v Bulgaria the Committee pointed out that: The burden of proof should not rest entirely on the complainant, but should be the subject of an appropriate adjustment. It also applies to the collective complaints procedure. The Committee therefore relies on the specific data sent to it by the complainant organisation, such as its statistics which show unexplained differences. It is then for the Government to demonstrate that there is no ground for this allegation of discrimination. 164

C. Recognising Statistics and Other Relevant Information as an Instrument to Trigger a Shift in the Burden of Proof in Proving Discrimination in Death Penalty Cases
For the individual complainant in discrimination cases it is, of course, very important to know how much evidence they need to provide to shift the burden of proof to the respondent. Is it simply enough to demonstrate that they have been treated less favourably than a person of a different race, sex, religion, sexual orientation, age or other legally protected category? Naturally, the required level of proof will, at least to some extent, be dependent on the facts of the specific case. 165 Some national case law, such as those developed by the British and Irish courts, have not recognised the mere difference in treatment and/or difference between the comparators as sufficient to trigger a shift in the burden of proof. 166 165 As noted by Alexander, showing that a particular capital prosecutor has sought the death sentence in a racially discriminatory manner will indeed provide for a more probative and persuasive claim of racial discrimination than producing the statistical evidence of state-wide significance; Alexander (n 46) 23. 166 Schiek, Waddington and Bell (n 8) 253. judges that the required criteria for shifting the burden of proof on the respondent are met. Indeed, one such instrument that could be used to trigger a shift in the burden of proof and has been accepted by domestic and international judicial and quasi-judicial bodies is the use of relevant statistics. The further question is, however, to what extent statistical data can help a claimant give rise to a presumption of discrimination, including in capital cases. 167 The collection of statistical data can be particularly relevant for various aspects of non-discrimination law, especially in the context of indirect discrimination claims where statistics can help establish disparate impact. Where a capital defendant has been treated less favourably than other persons with a different personal characteristic and statistics show that defendants with the same characteristics are systematically disadvantaged in capital cases, this can be sufficient to shift the burden of proof to the respondent to show that the less favourable treatment was unrelated to any discriminatory reasons. It is certainly very disturbing that in almost all death penalty jurisdictions the available statistics show clear gender disparities in death penalty impositions and executions. 168 Although, only on the basis of that data, one cannot speak about overt sexist prejudice, these figures in themselves may suggest a clear possibility of a widespread sexist culture in capital sentencing. Such a culture can have a potent impact on individual decision-makers in capital cases of which they themselves may be aware only faintly or not at all.
Usually, to justify an inference, a tribunal will first need to make findings of primary facts from which it will be legitimate to draw. If there are no such findings, then it is, of course, impossible to draw an inference. However, the sharp gender imbalance in the application of the death penalty displayed by relevant statistical information precludes the want of primary facts and suffices to entitle the tribunal to require the respondent to provide convincing evidence of non-sexist grounds. 169 In cases of indirect discrimination where the rules or practices at issue are neutral on the surface but disproportionately and unfavourably affect specific groups of persons by comparison to others in the same or similar situation, the statistical data can be particularly useful. Reference has already been made to a possible example of such indirect discrimination in capital sentencing processes, that is, if domestic legal provisions on death penalty procedures permit jurors or trial judges to take a capital defendant's employment status into consideration when deliberating whether to deliver a death sentence in a particular case. These legal rules are neutrally formulated but their even-handed application may result in far more Afro-Americans being sentenced to death due to the fact that Afro-Americans constitute a particularly high proportion of the unemployed population. Thus, the final outcome of such a judicial practice involves discriminatory implications of a racial nature. Production of statistical data showing that Afro-American people are a particularly disadvantaged group when it comes to the application of the death penalty because of the implementation of certain rules or practices can be sufficient to require the respondent state to provide a convincing alternative explanation for the existent figures.
Indeed, in the civil law context statistics showing persistent and striking racial or gender disparities would create a prima facie case necessary to shift the burden of proof from the claimant to the respondent. In the 167 In American pre-McCleskey jurisprudence it is possible to notice some judicial positions acknowledging the relevant role of social science research, including statistical analyses, in judicial decision-making processes in certain situations, such as that 'the inferences to be drawn from the statistics are for the factfinder, but the statistics are accepted to show the circumstances'; Gross (n 34) 1294. In discrimination cases, in particular, valid and reliable statistics can serve as important circumstantial evidence of discrimination. 168 Alexander also observes that men are sentenced to death disproportionately more than women in most states. He gives an example of Ohio where over 450 convicted killers have been executed since 1972, but only four have been women. However, he believes that unlike the Racial Justice Act, the enactment of a Gender Justice Act which would allow for showing sentencing disparities by statistical evidence would indirectly lead to abolition of the capital punishment in practice and therefore it is very unlikely that such legislation would be adopted by a certain legislator or that gender-based claims of discrimination in capital cases would be accepted by the courts; Alexander (n 46) 38. 169 Many of the arguments and critiques levelled against the use of statistical evidence in individual capital sentencing discrimination claims can be seen as over-generalised, implausible, misleading and far-fetched, including, for instance, poorly substantiated claims that statistical proof of discrimination is inherently inaccurate, unreliable and untrustworthy, and that recognising statistics as evidence sufficient to establish a prima facie case of discrimination would impose a burden on the respondent state that is too onerous and would render death penalty legislation ineffective and thus eventually lead to de facto abolition of the death penalty, or that such measure would result in significantly greater expenses imposed on taxpayers. Particularly misleading, if not absurd, are the arguments that the legal authorisation of the possibility to use statistical evaluations in individual capital cases would lead to racial or gender quotas (ie specified racial or gender proportions relating to offenders) in a given capital sentencing system. According to Baldus, Woodworth and Pulaski, statistical tests are only instrumental in determining whether differences in treatment of capital defendants/victims with different protected characteristics are, as a matter of fact, a product of unlawful discrimination between different groups and they have nothing to do with achieving certain quotas; Baldus, Woodworth and Pulaski (n 46) 402-403.
criminal justice context, including in discrimination litigation relating to capital cases, however, most US death penalty jurisdictions, including the federal death penalty legal system, tend to apply a substantially higher -in effect, nearly impossibly high, if not insurmountable -standard of proof, as observed by some authors. 170 The US constitutional provisions of the equal protection clause, for instance, require that a criminal defendant who complains of discriminatory decisions made during a criminal trial must show some evidence of both discriminatory effect and discriminatory intent. This implies that pertinent statistics credibly demonstrating individuals of a different race or gender from that of the complainant but who are otherwise similarly situated and were not prosecuted or selected for capital prosecution or sentenced to death will normally not suffice to establish a prima facie case of discrimination. Instead, such evidence of discriminatory effect must be combined with further evidence of discriminatory intent in order to shift the evidentiary burden from the claimant to the respondent. This cumulative approach constitutes an extremely high standard which, on its face, contradicts the standard of proof originating from both the International Convention on the Elimination of All Forms of Racial Discrimination and the EU non-discrimination directives, according to which it is enough to prove the discriminatory effect of a certain policy, provision or practice in order to trigger a shift in the burden of proof. 171 While domestic courts do not seem to require any strict threshold for statistical evidence, 172 both the European Court of Human Rights and the Court of Justice of the European Union pointed out that a clear, substantial and remarkable difference (in numbers or percentages) between disadvantaged groups and other groups must figure in undisputed official statistics in order to establish a prima facie indication or presumption that a certain rule, practice or measure is discriminatory in effect. 173 Where this is the case, it is then for the respondent government to show that such disproportion is the result of objective factors unrelated to any discrimination on the prohibited grounds. More interestingly, in the Seymour-Smith case, the Court of Justice of the European Union suggested that a lower level of disproportion could still prove indirect discrimination 'if it revealed a persistent and relatively constant disparity over a long period' between groups of different sexes. 174 As for other possible grounds of discrimination -aside from racial or ethnic origin and sex -there is very limited data collection in relation to capital cases in many death penalty jurisdictions. This implies that in cases of disparities related to religion or belief, sexual orientation, age and socio-economic status, capital defendants will be much less likely to invoke statistical evidence in order to trigger a shift in the burden of proof. However, it will not always be necessary to present statistical evidence in proving (indirect) discrimination in capital cases. On the whole, the need to support and prove a claim of discrimination with statistical data will be dependent on the facts of a particular case. Sometimes other available and reliable sources (such as reports, studies and assessments elaborated by reputable NGOs or the UN's treaty bodies monitoring human rights violations in states' parties) that support certain analyses will suffice to prove that a protected group is disproportionally affected. 175 The Inter-American Commission on Human Rights took a more reserved and cautious approach to the issue of statistical evidence in proving the discriminatory application of the death penalty, similar to that of the US Supreme Court elaborated in the famous McCleskey case. 176 In Celestine v United States the Inter-American Commission rejected the arguments put forward by the petitioner, namely that 'the results of statistical studies alone are evidence of invidious racial discrimination in the capital sentencing context'. 177 In the words of the Inter-American Commission, unlike in other decision-making processes, statistical evidence alone cannot be taken in capital sentencing as sufficient to establish ' a prima facie case to prove the allegations of racial discrimination and partiality in the imposition of the death penalty such as to shift the burden of proof to the United States Government'. 178 This line of reasoning suggests that a death-sentenced claimant arguing discrimination in the capital sentencing process and the imposition of the death penalty will need to provide further evidence of the involvement of unequal treatment in their specific case in order to convince a reviewing judge (or other competent authority) that requirements for shifting the burden of proof onto the government respondent are met. This is to say that the first step for the claimant will be to submit statistical evidence that justifies and upholds the prevalence of the death penalty being applied more frequently to members of one (minority) group than to those of another.
The second requirement will be much more difficult to fulfil since it requires the claimant to establish the facts from which it may be presumed that there has been direct or indirect discrimination in applying capital punishment in their particular case. This additional factor in proving discrimination in capital sentencing cases, which must be clearly demonstrated in order to shift the burden of proof from complainant to respondent, almost always employs either direct evidence, such as testimony or admissions, or circumstantial evidence, such as qualitative or statistical data, showing a prima facie case. Prof Streib is absolutely right in saying that it will be ' extremely hard' 179 to obtain admissions of racial, gender, religious and other such stereotypes and prejudice on the part of prosecutors, judges and jury members which implies that courts will be required to rely on the available circumstantial evidence. This is also why the decisionmaking process in capital cases should not be judged according to the heightened standard of proof which, in fact, is much stricter than the ordinary equal protection standards when it comes to claims of discriminatory decisions. Too often such an ample discretion on the part of decision-makers in capital cases opens the entire criminal process to the possibility of delivering arbitrary and biased decisions. 180

V. Conclusions and recommendations
The foregoing has shown that some death penalty jurisdictions, most notably, the majority of US capital sentencing legal systems, pursue inappropriate standards of proof when it comes to proving discrimination in capital cases since an individual capital defendant challenging discrimination is faced with an undue burden (both in terms of scope and resources) to prove the discriminatory intent on the part of the decisionmaker(s) in order to prevail in their case; this is a particularly high burden of proof which is practically impossible to meet. With these considerations and findings as a backdrop, the concept of shared burden of proof -as stipulated in the EU non-discrimination directives -which recognises discriminatory effect per se (especially in the case of indirect discrimination) as a sufficient element to establish a prima facie case of discrimination can be said to correspond much more closely to the contemporary patterns of pervasive, mostly hidden, unintentional, unconscious, but still harmful bias among decision-makers as potential discriminators, including in capital cases.
The EU anti-discrimination directives spell out that, in proving direct or indirect discrimination in discrimination litigation, the burden of proof shall be transferred from the claimant to the respondent once the presumption of unequal treatment on the grounds of protected characteristic(s) is successfully established before a court or other competent authority. Yet, the Inter-American Commission on Human Rights appears to prefer a much stricter standard of proof according to which the applicant/petitioner needs to provide overt, clear-cut and ample evidence showing discriminatory bias on the part of decision-makers in a particular capital case to recognise a prima facie case of discrimination. Such an approach which is most evident from the Commission's decisions in Celestine and Andrews can be criticised for several reasons. First, one needs to be fully aware of the fact that each death penalty proceeding is an extremely demanding decision making and governance process ultimately resulting in a decision on someone's life or death. Deciding criminal cases, particularly capital cases where the defendant's life is at stake, should therefore meet much higher non-discrimination standards than, say, deciding on who is offered a job at a certain company or who is promoted or degraded in the workplace, or who gets social housing, or who becomes a member of a given organisation, association or club. Since capital punishment is irrevocable by its nature, it requires a greater degree of scrutiny than other punishments. Therefore, as far as the allocation of the burden of proof between the capital criminal defendant and the respondent government in discrimination litigation is concerned the legal standard should draw its guidelines from European non-discrimination law, in particular from the basic minimum requirements contained in the core EU anti-discrimination directives, pursuant to which the initial burden would rest with the claimant (capital defendant) to present evidence that suggests, on the balance of probabilities, that the death sentence in their particular case resulted from a discriminatory decision making process. Accordingly, where the claimant has been successful in proving the facts from which conclusions could be drawn that one or more grounds of unlawful discrimination gave rise to a less favorable or disadvantageous decision influencing the outcome of a capital sentencing process, the burden of proof should move to the respondent. Put another way, the capital defendant alleging an impermissible discriminatory prosecutorial and/or judicial approach in their capital case should be asked merely to establish a presumption of discrimination and not be expected to carry the ultimate burden of persuasion concerning the existence of discrimination.
Second, the legal standard of proof to be applied in discrimination litigation in the capital sentencing context should take into consideration the real-life situations, challenges and obstacles confronted by claimants. Given that in everyday life the most overt and straightforward cases of discrimination can hardly ever be found, mainly because the decision-makers will seek to remove all traces and conceal relevant evidence of direct discrimination, the relevant standard of proof should definitely take into account the fact that discrimination in death penalty cases goes beyond obvious, intentional and purposeful examples of bias and prejudice. Therefore, the claimant should not be required to prove the existence of stereotypes/prejudice or discriminatory motive/intent on the part of decision-makers in capital cases.
In some states applying the death penalty -with the US at the forefront -it is thus imperative to amend their old-fashioned constitutional and legal provisions that require clear and convincing evidence of willful discrimination to be presented by the complainant in the capital sentencing context, 181 not least, because such a specific touchstone contravenes relevant international legal provisions, namely Article 1(1) of the International Convention on the Elimination of All Forms of Racial Discrimination, as well as Article 1 of the Convention on the Elimination of All Forms of Discrimination Against Women. Both Conventions provide for a broad definition of racial and gender discrimination, which make clear that States Parties to the convention must protect against conduct which produces a discriminatory effect, not merely against conduct that reveals a discriminatory intent. 182 In the same vein, the advanced and sophisticated European non-discrimination law predicated on Article 14 of the Convention for the Protection of Human Rights and Fundamental Freedoms, Protocol No. 12 to the Convention and, in particular, the EU non-discrimination directives and further explored and interpreted by two important judicial institutions, the European Court of Human Rights and the Court of Justice of the European Union, suggests that the (discriminatory) motive behind less favourable treatment is irrelevant, for it is the impact that is really important.
Third, capital punishment may be imposed only when the guilt of the person charged is based upon clear and convincing evidence leaving no room for a biased explanation of facts. Indeed, in reviewing various issues related to death penalty cases, including the allegations of the discriminatory decision-making process, a heightened level of judicial scrutiny shall be employed. Relevant human rights treaty bodies and institutions, such as the UN Human Rights Committee, the Inter-American Commission on Human 181 See also, in regard to necessary legislative and judicial reform, the proposition advanced by Ronald J Tabak that legislative bodies should be the first to address the problem of constitutional claims of discrimination in death sentencing and thereby eventually compel the courts to do so as well. Ronald J Tabak, 'Is racism irrelevant? Or should the fairness in death sentencing act be enacted to substantially diminish racial discrimination in capital sentencing? ' (1990-1991)  Rights, and the UN Special Rapporteur on extrajudicial, summary, or arbitrary executions, reiterated on several occasions that all safeguards and guarantees for a fair trial enshrined in pertinent international legal instruments must be fully respected and ensured when it comes to criminal cases possibly leading to the imposition of capital punishment. Death penalty jurisdictions have an enhanced obligation to ensure that any deprivation of life which may occur through the application of the death penalty complies strictly with the requirements of the applicable international human rights instruments. Such a restrictive approach, which requires a 'heightened scrutiny test', was developed by international human rights authorities, including the Inter-American Commission on Human Rights. This means that capital cases must always meet the highest and exacting standards of independence, competence, objectivity and impartiality of judges and jurors. Those accused of capital crimes are, in accordance with international human rights law, entitled to a special and enhanced protection with regard to their due process rights, the protection of which is above and beyond that afforded in non-capital cases. If there is, for example, a serious doubt or presumption that a capital sentencing process was tainted with biased and prejudiced treatment of a defendant and/or a discriminatory decision, the death sentence must be quashed and a fair retrial shall be provided. The new sentencing hearing must be conducted in accordance with the equality, due process and fair trial protections and must include also competent legal representation.
Finally, and perhaps more fundamentally, the petitioner argued in Celestine that 'the present US rule of law requiring the defendant to prove racial discrimination in his trial is an unrealistic standard of review because no capital defendant has ever met that burden' and that 'subtle, system-wide racial discrimination is most often evident only through large statistical studies'. 183 He also maintained that 'if reliable statistical studies demonstrate the likelihood of racial discrimination within the criminal justice system, the burden must shift to the Government to prove that the capital hearing was free of racial discrimination'. 184 Moreover, in the same case the petitioner also correctly pointed out that a prohibition on racial discrimination constitutes a peremptory norm of general international law (jus cogens) which 'places "a heavy burden of justification upon the United States for the continuation of existing legal doctrines and policies that have permitted this state of affairs"'. 185 Regrettably, the Inter-American Commission on Human Rights dismissed these rather cogent arguments as not persuasive enough and held that the statistics provided were not sufficient to make a prima facie case that the death sentence was the result of racial discrimination so as to shift the burden of proof to the US government. Here it is necessary to indicate that, as the foregoing has demonstrated, comprehensive statistical evaluations of racial, ethnic, sexual, religious and other disparities, especially those made in the form of multiple-regression analysis and by virtue of official statistics, or other such relevant sources of information, when showing an outstanding and clear pattern of less favourable or disadvantageous effects towards specific group(s) of criminal defendants/victims of a capital crime and where there is no legitimate justification for such disparities, should carry the same evidentiary value and probative force as in other spheres of an individual's life and work. Consequently such statistical evaluation should normally be per se sufficient to establish the presumption of discrimination in the capital sentencing context, where a higher level of judicial scrutiny is required. 186 In any event, veracious statistical evidence showing a discriminatory effect can serve as additional corroborative evidence to shift the burden of proof to the respondent. 187 The court will then need to evaluate whether the complainant has demonstrated discrimination in light of the government's evidence. 188 The state must be given an opportunity to rebut the discrimination claim by explaining that legitimate and discriminatory ground-neutral criteria produced a disproportionate and discriminatory result. In order to discharge its burden, the alleged perpetrator must provide evidence that shows, on the balance of 183 Willie L. Celestine v United States (n 2) para 17. 184 ibid. 185 ibid para 18. 186 In the view of the Committee on the Elimination of Racial Discrimination, the relevant provisions of the International Convention on the Elimination of All Forms of Racial Discrimination prohibit -in terms of unlawful discriminatory effect -those race-neutral laws, actions and practices that have an 'unjustifiable disparate impact' on a particular racial or ethnic group. Committee on the Elimination of Racial Discrimination, General Recommendation No 14: Definition of discrimination (Art. 1, par. 1), 22 March 1993, A/48/18 at 114 (1994), para 2. 187 Thus, the use of statistics represents a tool to 'link individual direct discrimination cases to broader patterns of inequality'; Schiek, Waddington and Bell (n 8) 322. 188 For instance, Alexander also maintains that, despite the fact that statistics are imperfect, they should be accepted as a possible tool to help prove a prima facie case of discrimination, especially because of the extremely serious nature of capital punishment reflected in its complete finality and irreversibility; Alexander (n 46) 26.
probabilities, that the prohibited grounds of discrimination did not play any role in the given capital sentencing decision-making process and that all decisions leading to the death sentence were made in an absolutely fair, impartial and unbiased manner. In the absence of such evidence, the statistical evidence can be accepted as sufficient to demonstrate that personal factors entered into the decision that resulted in the defendant's death sentence. 189 In time, death penalty jurisdictions will need to pay heed to international human rights jurisprudence and case law elaborated by the human rights treaty bodies concerned if they do not want to find themselves in a vicious cycle of repeated international criticism for their responsibility for human rights violations and in order to get rid of increasing international community pressure to conform their death penalty systems to international standards. According to these standards, a great responsibility rests with the national judges in discharging their duties in capital trials. They are required to ensure a fair, impartial and non-biased sentencing process as well as a non-discriminatory decision on punishment in each particular capital case. This means, first of all, that they themselves must be of high principles and morality who permit no exceptions from international rules on fair and impartial trials and equal treatment before courts and tribunals. By the same token, they should also require the same criteria to be fulfilled by other prosecutorial and judicial staff and jury members involved in the capital sentencing cases. As soon as the reasonable suspicion of bias on the part of one or more jurors arises, the trial judge must carry out all necessary measures to ensure that the continuation of the criminal process is conducted in an equitable and just manner. This may include a disqualification of biased and prejudiced jurors if necessary.
At the appellate level, the role of the reviewing judge will remain equally important as they will need to establish whether the capital sentencing process and the conviction in particular, including the imposition of the death penalty, were influenced by certain stereotypes and prejudice on the part of the particular decision-maker(s) who decided the case in the first instance. Moreover, the Inter-American Commission on Human Rights emphasized in Andrews v United States and Ramos v United States that a great responsibility in capital cases lies with the trial judges to ensure that the whole trial process is conducted in an impartial, fair and non-discriminatory manner. 190 Whenever there is a reasonable suspicion that a jury made a biased and discriminatory decision about the capital defendant's criminal liability and/or punishment, trial judges must take all necessary steps, including adequate examination and verification of serious allegations and supportive evidence submitted by the legal defence, and questioning and replacing jury members if necessary, 191 with a view to satisfy the requirements of equality before the law as guaranteed by Article II of the American Declaration of the Rights and Duties of Man and other relevant international human rights instruments.