Blind Justice: Why the Court Refused to Accept Statistical Evidence of Discriminatory Purpose in McCleskey v. Kemp—And Some Pathways for Change

Opening Address

BLIND JUSTICE: WHY THE COURT REFUSED TO

ACCEPT STATISTICAL EVIDENCE OF DISCRIMINATORY PURPOSE IN MCCLESKEY V. KEMP—AND SOME PATHWAYS FOR CHANGE

Reva B. Siegel

ABSTRACT—In McCleskey v. Kemp, the Supreme Court refused to accept statistical evidence of race discrimination in an equal protection challenge to the death penalty. This lecture, on the decision’s thirtieth anniversary, locates McCleskey in cases of the Burger and Rehnquist Courts that restrict proof of discriminatory purpose in terms that make it exceedingly difficult for minority plaintiffs successfully to assert equal protection claims.

The lecture’s aims are both critical and constructive. The historical reading I offer shows that portions of the opinion justify restrictions on evidence to protect prosecutorial discretion, while others limit proof of discrimination in ways that seem responsive to conservative claims of the era about race, rights, and courts. Scrutinizing the Court’s reasons for restricting inferences from statistical evidence opens conversations about the principles on which McCleskey rests and the decision’s prospective reach.

A close reading of the decision has led some courts to interpret McCleskey’s restrictions on statistical evidence as a response to particular concerns raised by the record in that case, opening the door to statistical evidence of bias in other equal protection challenges in criminal cases. At the same time, revisiting McCleskey and its progeny raises questions about the capacity of courts to redress bias in the criminal justice system. Three decades of living with McCleskey teaches that it is important to design remedies for bias in the criminal justice system that do not depend solely on judges for their implementation.

AUTHOR—Nicholas deB. Katzenbach Professor of Law, Yale University. This talk was given as the opening address of the Northwestern University Law Review’s 2017 Symposium: “A Fear of Too Much Justice”? Equal Protection and the Social Sciences 30 Years After McCleskey v. Kemp, a symposium held at Northwestern Pritzker School of Law, Oct. 20, 2017, [https://perma.cc/ http://www.northwesternlawreview.org/symposium

JTX8-DTAY]. I have titled my talk to mark the occasion of the thirtieth anniversary of McCleskey (1987) and the fiftieth anniversary of Arlo Guthrie’s song Alice’s Restaurant (1967).

For comments on the manuscript, I thank Linda Greenhouse and Robert Post. For research assistance and wide-ranging conversation in the course of preparing these remarks, I thank Ali Cooper-Ponte, Dylan Cowit, Vinay Nayak, Harrison Stark, Leigh Terry, and Alisa Tiwari.

Fifty years ago in Alice’s Restaurant,1 Arlo Guthrie sang of “a typical case of American blind justice.”2 Thirty years ago in McCleskey v. Kemp,3 the Supreme Court refused to accept statistical evidence of discrimination in an equal protection challenge to the death penalty. The Justices have crafted equal protection case law so that it is hard for minority plaintiffs to challenge facially neutral state action—first requiring plaintiffs to prove discriminatory purpose and then making discriminatory purpose exceedingly difficult to prove.4 After decades of discouraging claimants like Warren McCleskey, the Court now has an equal protection docket dominated by majority plaintiffs challenging race-conscious civil rights remedies.5

At a time when incarceration rates were skyrocketing, the Supreme Court restricted the use of statistical evidence to prove discriminatory purpose in death penalty cases.6 Three decades later, with race still a virulent

force in American politics,7 there is renewed interest in discriminatory purpose doctrine.8 McCleskey’s history helps us think about the doctrine’s structure and how it could yet evolve, even as that history cautions against relying solely on courts for redress of bias in the criminal justice system.9

Thirty years ago in McCleskey v. Kemp,10 the Court was presented with the “Baldus study,”11 social science evidence showing that race—the race of the defendant and the race of the victim—played a significant role in determining who was selected for the death penalty in Georgia. McCleskey’s lawyers offered this evidence as part of a discriminatory purpose claim of race discrimination.12 The Court assumed the validity of the social science evidence showing that race mattered in determining whether a convicted felon received the death penalty but refused to consider it as evidence of discriminatory purpose in McCleskey’s case, suggesting that statistical evidence could not be used to prove discriminatory purpose in a death penalty case in the ways it was used in employment discrimination and other constitutional cases.13

Why did the Court narrow discriminatory purpose claims as it did in McCleskey? Beginning in the 1970s, the Burger Court had begun to shift the

work of civil rights reform from the courts to the political branches.14 The Rehnquist Court continued this trend in McCleskey. The Court explained that restricting the use of statistical evidence to prove discriminatory purpose in a death penalty case was needed to protect the exercise of prosecutorial discretion, yet said more, questioning whether discrimination explained disparities in the criminal justice system and asserting the importance of judges deferring to community controls on the criminal law.15

Reading McCleskey in its historical context has both critical and constructive purposes. The historical reading allows us to explore the very different kinds of considerations that animated the Court’s decision to reject statistical evidence of discriminatory purpose in the case, and to debate anew the prospective reach of the Court’s judgment. A close, critical reading of the case has led some courts to interpret McCleskey’s restrictions on statistical evidence as concerned only with protecting prosecutorial discretion, and to identify openings for statistical evidence of bias in the criminal justice system today.16 At the same time, examining McCleskey’s legacy raises deep questions about the capacity of courts to redress bias in the criminal justice system. It reminds us that it is important to design remedies for bias in the criminal justice system that do not depend solely on judges for implementation.17

To begin the day’s conversation, Part I offers a quick sketch of the Baldus findings and of McCleskey’s equal protection claim. Part II shows how the Rehnquist Court narrowed discriminatory purpose doctrine as it rejected McCleskey’s claim, on terms designed to restrict statistical evidence in proving discriminatory purpose claims involving the death penalty and possibly other aspects of the criminal justice system. Part III locates the Court’s hostility to statistical proof of discrimination in conservative mobilizations of the 1980s and explores the understanding of social stratification on which hostility to statistical proof of discrimination was premised. I conclude with a key implication of this reading of McCleskey for the critical social sciences: Employ pluricentric rather than juricentric interventions to combat racial bias in the criminal justice system.

Warren McCleskey was black. He killed a white police officer during a robbery. “The jury found two aggravating circumstances—the victim was a

police officer and the homicide occurred in the course of an armed robbery— and returned a death sentence.”18 In challenging his death sentence under the Eighth and Fourteenth Amendments, McCleskey’s lawyers introduced evidence that race was another unnamed “aggravating factor” in his sentencing decision. In earlier cases, lawyers had tried and failed to establish racial bias in sentencing through social science evidence.19

But McCleskey’s lawyers thought they had a study that was sufficiently particularized to show bias in the administration of the Georgia death penalty statute and, by extension, in McCleskey’s case. The Baldus study showed that race mattered in death sentencing in Georgia in two ways: death sentences were more likely in cases involving white victims and in cases involving black defendants. The Baldus study analyzed the murder cases of more than 2,000 defendants in Georgia during the 1970s.20 Looking at the race of the victim alone, “defendants charged with killing white persons received the death penalty in 11% of the cases, but defendants charged with killing blacks received the death penalty in only 1% of the cases.”21 David Baldus examined “230 variables that could have explained the disparities on nonracial grounds.”22 In one model, “even after taking account of 39 nonracial variables, defendants charged with killing white victims were 4.3 times as likely to receive a death sentence as defendants charged with killing blacks.”23

The Baldus study illuminated the hidden role of race in McCleskey’s sentencing. Baldus’s 230-variable model ranked death-eligible cases according to the offense’s level of aggravation.24 Baldus found that “the effects of racial bias were most striking in the midrange cases” of aggravation where the decisionmakers experienced the most discretion about whether to give the death penalty.25 Focusing on these cases, Baldus showed that “14.4% of the black-victim midrange cases received the death penalty, and 34.4% of the white-victim [midrange] cases received the death

penalty.”26 On Baldus’s analysis of the facts, McCleskey’s case fell “within the midrange.”27 And the study further showed that the race of the defendant mattered, though not as dramatically as the race of the victim.28 Thus, the Baldus study indicated that “black defendants, such as McCleskey, who kill white victims have the greatest likelihood of receiving the death penalty.”29 McCleskey’s lawyers—Jack Boger, Tony Amsterdam, and Deval Patrick among others30—set out to translate these empirical findings about capital sentencing in Georgia into an equal protection challenge to McCleskey’s death sentence.31 Invoking then-prevailing equal protection frameworks, they argued that the defendant in a capital sentencing case had the burden of proving purposeful discrimination, which the defendant in a capital sentencing proceeding could meet, quoting Washington v. Davis,32 “by showing that the totality of relevant facts gives rise to an inference of discriminatory purpose.”33 Once McCleskey met that burden, “the burden shifts to the State to explain adequately the racial exclusion.”34

But—and here we face the $64,000 question—did the evidence of the Baldus study “give rise to an inference of discriminatory purpose” within the meaning of the Court’s equal protection cases circa 1986–1987? McCleskey’s lawyers saw clear support in Castaneda v. Partida,35 a 1977 equal protection case decided the year after Davis, in which the Court had allowed Mexican-Americans to establish a prima facie case of intentional discrimination through statistics showing an underrepresentation of their group in grand jury selection.36 And McCleskey’s lawyers pointed to Village of Arlington Heights v. Metropolitan Housing Development Corp.,37 also decided the year after Davis, arguing that Arlington Heights allowed plaintiffs to show purpose, and here I quote their brief, by demonstrating “(i) the racial impact of the challenged action, (ii) the existence of a system affording substantial state discretion, and (iii) a history of prior

discrimination.”38 McCleskey’s lawyers also reminded the Court of its recent Title VII decision in Bazemore v. Friday39 addressing the use of multiple regression analyses in a Title VII pattern-and-practice case of wage discrimination against North Carolina public employees. “Reduced to its essence,” McCleskey’s lawyers argued:

Of course, distinguish is exactly what the Supreme Court’s decision in McCleskey’s case did. Almost eerily as if in reply to McCleskey’s lawyers, Justice Powell emphasized that statistical evidence that might suffice in those cases would not suffice to prove discriminatory purpose in McCleskey’s case. Justice Powell objected to McCleskey relying solely on statistical evidence to prove discriminatory purpose in his case.41 Yet he went further, and expressed wide-ranging skepticism about the use of statistical evidence to prove discriminatory purpose in any equal protection challenge in a criminal case. Even as the majority accepted the statistical “validity” of the Baldus study, it rejected statistics as an evidentiary basis for McCleskey’s equal protection claim:

While the Court had accepted statistical studies employing multiple regression analyses as evidence of discrimination in employment discrimination and equal protection cases such as Bazemore and Castaneda, the Court held that McCleskey’s case was different: “But the nature of the capital sentencing decision, and the relationship of the statistics to that decision, are fundamentally different from the corresponding elements in the

venire-selection or Title VII cases.”43 To erect a firewall between the criminal justice setting and those cases where the Court had accepted statistical evidence as raising inferences about discriminatory bias, the Court pointed to differences in institutional contexts. The Court emphasized that decisions in trial and sentencing contexts involve more variables than decisions in jury selection or Title VII cases—and involve discretion that is more important to protect.44 The Court also explained that, in other contexts, “the decisionmaker has an opportunity to explain the statistical disparity,” but in McCleskey’s case “the State has no practical opportunity to rebut the Baldus study” and “policy considerations” cautioned against “requiring prosecutors to defend their decisions to seek death penalties, often years after they were made.”45

While the Court emphasized differences in institutional contexts, it never once addressed what these institutions shared in common—the need to make decisions about persons of different races. After rejecting the Baldus study as insufficient proof of discriminatory purpose in McCleskey’s case, the Court seemed wholly uninterested in inviting other plaintiffs to explore what the “statistically valid”46 Baldus study or other statistical evidence might show about the risk of racial bias in capital sentencing or the criminal justice system more generally.47

The Supreme Court went on to reject McCleskey’s argument that the state had given juries and prosecutors broad discretion in sentencing that was susceptible to racially discriminatory abuse. “McCleskey . . . appears to argue that the State has violated the Equal Protection Clause by adopting the capital punishment statute and allowing it to remain in force despite its allegedly discriminatory application.”48 Citing Personnel Administrator of Massachusetts v. Feeney49 (and Wayte v. United States50), Justice Powell argued,

In short, “McCleskey would have to prove that the Georgia Legislature enacted or maintained the death penalty statute because of an anticipated racially discriminatory effect.”52 Justice Powell dismissed this possibility in a sentence, there being, on this definition of discriminatory purpose, “no evidence . . . that the Georgia Legislature enacted the capital punishment statute to further a racially discriminatory purpose.”53

Why were McCleskey’s lawyers unable to mobilize social science evidence of racial bias in the administration of Georgia’s death penalty into a successful equal protection challenge? The standard story points to the Court’s decision in Washington v. Davis barring disparate impact claims in constitutional cases as shielding the operations of the criminal justice system from equal protection challenge. McCleskey’s history reminds us just how much of this work is performed by subsequent cases that restrict how discriminatory purpose can be proved.

The standard story goes something like this: During the 1970s, President Richard Nixon appointed Justices to the Supreme Court54 who took important steps toward ending the Second Reconstruction when they rejected disparate impact claims and required plaintiffs alleging equal protection violations to prove discriminatory purpose in Washington v. Davis. As a student of Owen Fiss who graduated at the time of McCleskey, I was raised protesting Davis: I read Owen Fiss’s Groups and the Equal Protection Clause,55 Alan Freeman’s Legitimizing Racial Discrimination Through Antidiscrimination Law,56 and Chuck Lawrence’s The Id, the Ego, and Equal

Protection: Reckoning with Unconscious Racism (published at the time of McCleskey),57 and tried my own hand at a Davis critique a decade later, first in Rule of Love58 and then in Why Equal Protection No Longer Protects.59

The standard story is correct, to this extent: The Burger Court reversed decisions recognizing equal protection claims of unjustified racial disparate impact in 1976 in Davis and a year later in Arlington Heights. But these Burger Court cases did not in fact end the role of impact evidence in equal protection cases. The Burger and Rehnquist Courts did this work in a series of later decided cases that restrict the role of impact evidence in proving discriminatory purpose60—cases that include McCleskey.

Before the Supreme Court decided Davis in 1976—and for some years after—plaintiffs pointed to evidence of racial impact as evidence from which one could infer discriminatory purpose.61 As Justice John Paul Stevens explained in his concurring opinion in Davis: “Frequently the most probative evidence of intent will be objective evidence of what actually happened rather than evidence describing the subjective state of mind of the actor. For normally the actor is presumed to have intended the natural consequences of his deeds.”62

In Davis’s wake, the Court decided cases like Castaneda v. Partida, allowing plaintiffs to make out a prima facie case of intentional discrimination by statistical evidence,63 and Arlington Heights, allowing the introduction of disparate impact evidence and other circumstantial evidence from which inferences about discrimination might be drawn.64 Both these cases, decided the year after Davis, are cases on which McCleskey’s lawyers drew to build their statistical evidence into a prima facie case of discrimination.65

But by the end of the 1970s, the Burger Court had begun to restrict the role of impact evidence in proving discriminatory purpose.66 Feeney was key. In the 1979 sex discrimination decision Personnel Administrator of Massachusetts v. Feeney,67 the Burger Court held that discriminatory purpose can be found only if “the decisionmaker . . . selected or reaffirmed a particular course of action at least in part ‘because of,’ not merely ‘in spite of,’ its adverse effects upon an identifiable group.”68 By requiring a specific intent to harm, Feeney restricted the inferences about purpose that can be drawn from impact evidence.

In Columbus Board of Education v. Penick, a school desegregation decision decided just after Feeney, Justice Powell explained that Feeney had narrowly defined discriminatory purpose in order to limit the power of judges: “Courts, of course, should confront discrimination wherever it is found to exist. But they should recognize limitations on judicial action inherent in our system . . . .”69 Justice Powell urged the Court to adhere to the Feeney standard.70 Justice Powell again invoked Feeney to impose exacting standards for proof of claims of selective prosecution in his 1985 decision for the Court in Wayte v. United States.71

Like his opinions in Penick and Wayte, Justice Powell’s opinion in McCleskey discussed policy reasons for restricting proof of discriminatory

purpose. In McCleskey, Justice Powell expressed concern about protecting “discretion . . . essential to the criminal justice process.”72

With this, the Court restricted the inferences about discriminatory purpose that can be drawn from statistical evidence and separated proof of discriminatory purpose in death penalty (and possibly other criminal) cases from proof of discrimination in equal protection-jury cases and Title VII cases.

In restricting the use of statistics to prove discriminatory purpose, the Court emphasized its concern about protecting the discretion of juries and prosecutors. In part the Court seemed to suggest that it was acceptable to protect discretion fundamental to the process in this way because of the safeguards against race discrimination already built into the system.74 But long concluding passages of the opinion sounded a very different note, suggesting that restricting the use of statistics was necessary to protect the criminal justice system against discrimination claims and to protect federal courts from entanglement in politically fraught questions. These sections of the opinion introduce reasons for restricting proof of discriminatory purpose not tied to protection of prosecutorial discretion, and are considerably more troubling.

Writing for the Court, Justice Powell recognized that “if we accepted McCleskey’s claim . . . we could soon be faced with similar claims as to other types of penalty” and he foresaw claims based on “unexplained discrepancies that correlate to membership in other minority groups, and even to gender.”75 The Court understood that statistical challenges would not

be cabined to death, or to race, because there were many unexplained disparities in the American criminal justice system that raise inferences of bias. To prevent these claims from ever materializing (“there is no limiting principle to the type of challenge brought by McCleskey”), the McCleskey majority announced a constitutional principle of indeterminate normative significance and practical scope: “The Constitution does not require that a State eliminate any demonstrable disparity that correlates with a potentially irrelevant factor in order to operate a criminal justice system that includes capital punishment.”76

Justice William Brennan famously termed this concern “a fear of too much justice.”77 But what kind of fear is that? On what normative ground does the majority base this constitutional principle limiting equal protection claims? What exactly is the majority saying in concluding the McCleskey opinion as it does?

In voicing skepticism about statistical disparities, the Court was no longer discussing prosecutorial discretion but seemed instead to be engaging in contemporary debates about discrimination. At the time of McCleskey, conservatives in the Reagan administration and Meese Justice Department objected that the law was too quick to draw inferences that discrimination explained statistical disparities. They criticized statistical proof of discrimination in disparate treatment—pattern-and-practice cases, disparate impact, and affirmative action cases of the day—EEOC v. Sears, Roebuck & Co.;78 Wards Cove Packing Co. v. Atonio;79 and City of Richmond v. J.A. Croson Co.80 to name a few. For example, in 1986, as Chair of the EEOC, Clarence Thomas famously attacked the EEOC’s pattern-and-practice case of sex discrimination against Sears because the case was based purely on statistics.81

Conservatives who criticized statistical measures as a means of drawing inferences about discrimination often expressed the belief that social stratification was better explained by legitimate rather than illegitimate factors.82 Even when statisticians carefully controlled for variables (David Baldus examined “230 variables that could have explained the disparities on nonracial grounds”83), critics were not satisfied. They believed statistical disparities were more likely caused by legitimate factors such as group differences in taste and talent than by illegitimate factors such as race discrimination, and for this reason believed that basing antidiscrimination liability on statistical evidence amounted to social engineering, an illegitimate government interference with naturally occurring inequalities among groups. An attack on disparate impact published by Reagan’s Office of Legal Policy the same year as McCleskey reasoned that disparate impact standards would alter “naturally occurring statistical disparities between groups that are [otherwise] inevitable in a heterogeneous society such as the United States” and, by doing so, lead to “the permanent institutionalization of race- and gender-conscious affirmative action.”84 A 1986 Harvard Law Review article by then-Commissioner on the United States Commission on Civil Rights Morris Abram similarly argued:

(As I have elsewhere pointed out, this belief in racial group difference might not be especially surprising were it not espoused by advocates of “colorblindness.”86)

The attack on statistical proof of discrimination was not just New Right music in the air. Conservative critics imported these concerns about group differences into the argument over the use of statistics in proving discrimination in the criminal justice setting. An amicus brief by the Washington Legal Foundation attacked McCleskey’s arguments as calling for “racial proportionality” in the administration of criminal justice.87 “Acceptance of petitioner’s argument would open the door to Title VII-style ‘disparate impact’ challenges to criminal sentences of all kinds.”88 “No workable system of criminal justice could accommodate the demands for race- and class-based parity in sentencing advanced by petitioner. Nor does the Constitution require a regime of ‘statistical justice’ which would subject the validity of every criminal sentence to the vagaries and manipulations of fluctuating demographic data.”89 It is “the considered judgment of our law that seemingly ‘disproportionate’ outcomes in terms of race or other characteristics are generally explainable by a host of legitimate factors other than actionable discrimination . . . .”90

These objections to statistical evidence of discrimination appear in McCleskey itself, in the passages of the opinion which worried that acceptance of McCleskey’s claim would invite constitutional challenges to all manner of criminal penalties based on “statistical disparities” and “unexplained discrepancies” relating to different demographic grounds, not only as to defendants, but as to all participants in the criminal justice system.91

It was to limit such claims that the opinion closed by restricting the inferences of bias that could be drawn from statistical discrepancies in a capital case: “The Constitution does not require that a State eliminate any demonstrable disparity that correlates with a potentially irrelevant factor in order to operate a criminal justice system that includes capital punishment.”92 Justice Powell offered no justification for this “fear of too much justice”93 claim other than to quote a fragment of Gregg v. Georgia94 for the newly constructed principle that “the Constitution does not ‘plac[e] totally unrealistic conditions on its use.’”95

But to what kind of realism is the Court appealing? What is the empirical or normative basis of this constitutional principle limiting statistical evidence of discrimination? In its closing, cryptic appeal to realism, the Court may have intimated doubt that racial discrimination accounted for disparities in the commission of crimes and adverted instead to explanations tied to racial group differences favored by conservatives.96

The Court’s appeal to realism surely warned of political limits on judges’ capacity to intervene in the criminal law. (McCleskey’s closing citation to Gregg adverted to the Court’s failed effort to impose constitutional limitations on the death penalty.97)

Justice Powell concluded the McCleskey opinion by insisting that the racial stratification of the criminal justice system, like the racial segregation of the public schools, was a political and not a legal question, best handled by legislatures and not by courts. The reasons he offered for discounting inferences about discrimination that could be drawn from statistical evidence in this passage of the opinion did not concern prosecutorial discretion but instead concerned “the moral values of the people.”98 Echoing Justice Byron White in Washington v. Davis,99 and his own opinion in Penick,100 Justice Powell asserted:

At a time when the political branches were engaged in a “War on Crime,”102 incarceration rates were skyrocketing,103 and racial discrepancies

in incarceration rates were in the headlines,104 the Court rejected McCleskey’s claim on the grounds that it was better suited for political than legal resolution. The Court identified problems with relying on statistics to prove discriminatory purpose in the criminal justice system—while saying nothing about the costs to the constitutional order of excluding evidence of this kind.

* * *

For decades political liberals have criticized Washington v. Davis and complained that the Supreme Court required equal protection complainants to prove discriminatory purpose. But having required proof of purpose, perhaps as remarkable are the obstacles to proof of purpose that the Court has imposed.

No plaintiff can meet Feeney’s evidentiary burden when courts have chosen to apply it.105 And as numerous commentators have observed,

language in the McCleskey opinion offers another powerful obstacle to discriminatory purpose claims—allowing, if not encouraging, judges to refuse inferences of discriminatory purpose from statistical disparities, unless litigants are creative in coupling statistical and nonstatistical evidence as the plaintiffs were in Floyd v. City of New York,106 New York’s stop-andfrisk case. At the time of McCleskey, Randy Kennedy107 and Skip Gates108 described the limits on statistical proof of purpose that the decision helped to consolidate and enforce. Decades later, Aziz Huq,109 Michael Selmi,110 and David Sklansky among others111 report much the same story. interest

in discriminatory purpose doctrine resurgent,112 there is good reason to examine the constraints on proof of

With racism, and

purpose that the Rehnquist Court incorporated into equal protection law as courts were ending the Second Reconstruction. It may be that these constraints on the kinds of evidence that judges may consider in evaluating claims of discriminatory purpose are due for reconsideration.

Although Justice Powell did not write McCleskey in such a way as to clarify the conditions warranting restrictions on the use of statistical evidence, judges can read McCleskey with attention to these questions. For instance, some courts have implied that they might consider statistical studies that are coupled with circumstantial evidence113 or that focus on the specific prosecutorial officer or entity at issue.114 In cases not involving the death penalty or prosecutorial discretion, courts consider statistical evidence of discriminatory policing without pausing to distinguish McCleskey.115

The use of statistics in these cases present openings for change. Judges may not be moved by the same considerations that seemed to guide the McCleskey Court, especially when judges can see the effects of decades of judicial abdication. Still, the spirit of McCleskey continues to guide many judges.

I am a student of Owen Fiss’s. I do not give up on courts.116 But I am a student of Owen Fiss’s who graduated in the era of Bowers v. Hardwick117 and McCleskey. McCleskey was decided by a Court openly skeptical about its role in enforcing equality rights—a Court of the kind the nation has long faced and in all likelihood will face in the near term. For this reason, as well as many others, advocates interested in securing enforcement of the Constitution’s equality guarantees in the criminal justice system cannot rely on courts alone. We need pluricentric—rather than juricentric—modes of enforcing the Constitution.118

Empirical social science can illuminate bias in the criminal justice system. But given an unwilling—or even reticent—judiciary, it may not make best sense to imagine judges as the primary agents for implementing the insights of critical social science, as, for example, much writing on

implicit bias has.119 Indeed, the best understanding of the social science may counsel putting it into practice outside courtroom settings.120

For these and other reasons, scholars have channeled implicit bias interventions directly into the criminal justice system, where the research can help decisionmakers on the street and on the bench become aware of the unconscious biases they may have.121 Actors in representative branches of

government can require law enforcement personnel to participate in debiasing trainings, just as legislators may require themselves to take account of the racial impact of their decisions122—even if judges never interpret the Constitution to require it.

None of these interventions takes the place of a state or federal judge constraining discriminatory excesses in the criminal justice system. But each intervention illustrates biases that exist in the criminal justice system, and demonstrates social-scientific tools available to identify and begin to address them. These interventions can make a crucial difference in their own right. And they can establish roles for social-scientific tools in the criminal justice system so that courts might over time come to see reasons to allow and even to mandate their use.

--- FOOTNOTES ---

Copyright 2018 by Reva B. Siegel

Printed in U.S.A. Vol. 112, No. 6

INTRODUCTION ..................................................................................................... 1270 I. MCCLESKEY’S EQUAL PROTECTION ARGUMENT ............................................... 1272 II. MCCLESKEY AND DISCRIMINATORY PURPOSE AS A DOCTRINE OF

JUDICIAL DEFERENCE .................................................................................... 1277

III. HOSTILITY TO STATISTICAL PROOF OF DISCRIMINATION AND REASONING

ABOUT RACE IN MCCLESKEY .......................................................................... 1280 CONCLUSION: SECURING EQUAL TREATMENT, WITH OR WITHOUT COURTS .............. 1289

INTRODUCTION

1 ARLO GUTHRIE, Alice’s Restaurant Massacree, on ALICE’S RESTAURANT (Warner Bros. 1967). 2 Id. (“Obie came to the realization that it was a typical case of American blind justice, and there

wasn’t nothing he could do about it . . . .”).

3 481 U.S. 279 (1987). 4 See Reva B. Siegel, The Supreme Court, 2012 Term—Foreword: Equality Divided, 127 HARV. L. REV. 1, 9–23 (2013) [hereinafter Siegel, Equality Divided]; see also Reva Siegel, Why Equal Protection No Longer Protects: The Evolving Forms of Status-Enforcing State Action, 49 STAN. L. REV. 1111, 1113 (1997) [hereinafter Siegel, Why Equal Protection No Longer Protects].

5 See, e.g., Siegel, Equality Divided, supra note 4, at 1–58. 6 For McCleskey’s social context, see infra notes 102–104 and accompanying text. For commentary

on McCleskey’s impact in the legal system, see infra notes 107–111 and accompanying text.

7 The Supreme Court has recently suggested that race discrimination is waning. See Shelby County v. Holder, 133 S. Ct. 2612, 2618 (2013) (“There is no denying . . . that the conditions that originally justified these measures no longer characterize voting in the covered jurisdictions.”). But there is abundant evidence to the contrary. See, e.g., First Amended Complaint at 2, Sines v. Kessler, No. 3:17cv-00072-NKM (W.D. Va. Jan. 5, 2018) (“Over the weekend of August 11 and 12, 2017, hundreds of neo-Nazis and white supremacists traveled from near and far to descend upon the college town of Charlottesville, Virginia . . . .”).

Intent,

130 HARV.

8 Sam Bagenstos has criticized scholars’ preoccupation with implicit bias in part because it “suggests that the most prevalent form of discrimination is unconscious, when conscious bias, stereotyping, and prejudice may be just as important in practice.” See Samuel R. Bagenstos, Implicit Bias’s Failure, 39 BERKELEY J. EMP. & LAB. L. (forthcoming 2018). Other scholars have called for renewed engagement with discriminatory purpose doctrine. See, e.g., Richard H. Fallon, Jr., Constitutionally Forbidden Legislative that REV. “intent-, purpose-, and motivation-based doctrines cry out for reexamination and reform”); Aziz Z. Huq, Judging Discriminatory Intent, 103 CORNELL L. REV. (forthcoming 2018) (manuscript at 6) (on file with the Northwestern University Law Review) (showing that four cases “all from a single four-month period in 2017—show the problem of discriminatory intent to be pervasive” but that “forty or so years’ applying a ‘discriminatory intent’ touchstone have not elicited either a stable formulation of the term or a predictable evidentiary framework for identifying such intent”); Michael Selmi, Statistical Inequality and Intentional (Not Implicit) Discrimination, 79 LAW & CONTEMP. PROBS. 199, 202 (2016) (arguing that “observed disparities should not be confused with implicit bias as repeated patterns of behavior will almost certainly have a conscious component to them”).

(arguing

(2016)

523,

L.

9 See infra notes 116–122 and accompanying text. 10 (1987). 11 Id. at 286. 12 Brief for Petitioner at 80–89, McCleskey, (No. 84-6811). The plaintiffs also raised

an Eighth Amendment challenge. Id. at 41.

13 See infra notes 41–46 and accompanying text.

I. MCCLESKEY’S EQUAL PROTECTION ARGUMENT

14 Siegel, Equality Divided, supra note 4, at 20–23; see infra Part II. 15 See infra Part III. 16 See infra notes 113–115 and accompanying text. 17 See infra notes 116–122 and accompanying text.

18 David C. Baldus et al., Reflections on the “Inevitability” of Racial Discrimination in Capital Sentencing and the “Impossibility” of Its Prevention, Detection, and Correction, 51 WASH. & LEE L. REV. 359, 364 (1994) (citing McCleskey v. Kemp, , 284–85 (1987 .

19 In 1968, the Eighth Circuit, in an opinion by then-Judge Harry Blackmun, rejected such social science evidence on the grounds that it was not sufficiently detailed and that it failed to show intentional discrimination in the defendant’s case. Maxwell v. Bishop, 398 F.2d 138, 146–47 (8th Cir. 1968), vacated, 398 U.S. 262 (1970); see Samuel R. Gross, David Baldus and the Legacy of McCleskey v. Kemp, 97 IOWA L. REV. 1905, 1906 (2012).

20 McCleskey, 481 U.S. at 286. 21 Id. 22 Id. at 287. 23 Id. 24 Id. at 287 n.5. 25 Id.

26 Id. 27 Id. 28 The McCleskey majority opinion reports that “black defendants were 1.1 times as likely to receive

a death sentence as other defendants.” Id. at 287.

29 Id. 30 See JEFFREY L. KIRCHMEIER, IMPRISONED BY THE PAST: WARREN MCCLESKEY AND THE

AMERICAN DEATH PENALTY 306 (2015).

31 For Jack Boger’s account of the case, see John Charles Boger, McCleskey v. Kemp: Field Notes

from 1977–1991, 112 NW. U. L. REV. 1637 (2018).

32 426 U.S. 229 (1976). 33 Brief for Petitioner, supra note 12, at 47 (quoting Davis, 426 U.S. at 239–42). 34 Id. at 47–48 (citing Alexander v. Louisiana, 405 U.S. 625, 632 (1972 . 35 430 U.S. 482 (1977). 36 Id. at 494–95 (noting that the burden then shifts to the state to rebut the case). 37 429 U.S. 252 (1977).

[P]etitioner’s submission to the Court is a simple one. Evidence of racial discrimination that would amply suffice if the stakes were a job promotion [Bazemore], or the selection of a jury [Castaneda], should not be disregarded when the stakes are life and death. Methods of proof and fact-finding accepted as necessary in every other area of law should not be jettisoned in this one.40

Our assumption that the Baldus study is statistically valid does not include the assumption that the study shows that racial considerations actually enter into any sentencing decisions in Georgia. Even a sophisticated multiple-regression analysis such as the Baldus study can only demonstrate a risk that the factor of race entered into some capital sentencing decisions and a necessarily lesser risk that race entered into any particular sentencing decision.42

38 Brief for Petitioner, supra note 12, at 50. 39 478 U.S. 385 (1986). 40 Brief for Petitioner, supra note 12, at 31–32. 41 McCleskey v. Kemp, , 292–93 (1987) (“Thus, to prevail under the Equal Protection Clause, McCleskey must prove that the decisionmakers in his case acted with discriminatory purpose. He offers no evidence specific to his own case that would support an inference that racial considerations played a part in his sentence. Instead, he relies solely on the Baldus study.”).

42 Id. at 291 n.7.

But ‘“[d]iscriminatory purpose’ . . . implies more than intent as volition or intent as awareness of consequences. It implies that the decisionmaker, in this case a state legislature, selected or reaffirmed a particular course of action at

43 Id. at 294. 44 Id. at 294–95. 45 Id. at 296 (internal quotation marks omitted). 46 Id. at 291 n.7 (“As did the Court of Appeals, we assume the study is valid statistically without

reviewing the factual findings of the District Court.”).

47 For examples, see infra notes 119–122 and accompanying text. 48 McCleskey, 481 U.S. at 297–98. 49 442 U.S. 256 (1979). 50 470 U.S. 598 (1985). Wayte is a case alleging selective prosecution for failure to register for the

military draft that incorporates Feeney standards. See infra note 71 and accompanying text.

least in part ‘because of,’ not merely ‘in spite of,’ its adverse effects upon an identifiable group.”51

II. MCCLESKEY AND DISCRIMINATORY PURPOSE AS A DOCTRINE OF JUDICIAL DEFERENCE

51 McCleskey, 481 U.S. at 298 (alteration in original) (quoting Feeney, 442 U.S. at 279) (citing

Wayte, 470 U.S. at 608–09).

52 Id. 53 Id. 54 President Nixon appointed Justices Warren Burger, Harry Blackmun, Lewis Powell, and William Rehnquist. See, e.g., MICHAEL J. GRAETZ & LINDA GREENHOUSE, THE BURGER COURT AND THE RISE OF THE JUDICIAL RIGHT 4–5 (2016) (describing how “Nixon got his four Supreme Court appointments. Warren Burger was the first [in May 1969] . . . followed by those of Harry Blackmun in May 1970 and Lewis F. Powell, Jr., and William H. Rehnquist in December 1971”); KEVIN J. MCMAHON, NIXON’S COURT 6 (2011) (noting that Rehnquist was Nixon’s “conservative” jurist for the Court, while Burger, Blackmun, and Powell were “his more moderate selections”).

55 Owen M. Fiss, Groups and the Equal Protection Clause, 5 PHIL. & PUB. AFF. 107 (1976). 56 Alan David Freeman, Legitimizing Racial Discrimination Through Antidiscrimination Law: A

Critical Review of Supreme Court Doctrine, 62 MINN. L. REV. 1049 (1978).

57 Charles R. Lawrence III, The Id, the Ego, and Equal Protection: Reckoning with Unconscious

Racism, 39 STAN. L. REV. 317 (1987).

58 Reva B. Siegel, “The Rule of Love”: Wife Beating as Prerogative and Privacy, 105 YALE L.J.

2117 (1996).

59 Siegel, Why Equal Protection No Longer Protects, supra note 4. 60 See, e.g., Siegel, Equality Divided, supra note 4, at 15–16. 61 See, e.g., id. at 16–17. 62 , 253 (1976) (Stevens, J., concurring). 63 430 U.S. 482, 500–01 (1977). 64 429 U.S. 252, 266 (1977). 65 Brief for Petitioner, supra note 12, at 52–53, 56 (first quoting Arlington Heights, 429 U.S. at 266; then quoting Castaneda, 430 U.S. at 494). Section II of this brief sets out a prima facie case of discriminatory purpose under Arlington Heights by first showing “The Racial Disparities” and observing “[t]he impact of the official action—whether it ‘bears more heavily on one race than another’ . . .— provides[s] an important starting point.” Id. at 52 (alterations in original) (citing Arlington Heights, 429 U.S. at 266). “Here, the Baldus studies reveal substantial, unadjusted racial disparities: a death-

sentencing rate nearly eleven times higher in white-victim cases than in black-victim cases.” Id. at 52– 53.

66 See, e.g., Siegel, Equality Divided, supra note 4, at 17–19. For another historical account of the evolution of discriminatory purpose doctrine in this era, see Ian Haney-López, Intentional Blindness, 87 N.Y.U. L. REV. 1779 (2012). 67 442 U.S. 256 (1979). 68 Id. at 279 (citing United Jewish Orgs. v. Carey, 430 U.S. 144, 179 (1977) (Stewart, J., concurring

in judgment .

69 443 U.S. 449, 489 (1979) (Powell, J., dissenting). 70 Id. at 510–11 (Rehnquist, J., dissenting). Justice Byron White appealed to concerns sounding in separation of powers and federalism as reasons for requiring discriminatory purpose doctrine in Davis. , 248 (1976).

71 470 U.S. 598 (1985). Wayte involved a claim of selective prosecution for evasion of the military draft. Justice Powell, writing for the majority, emphasized that “[i]n our criminal justice system, the Government retains ‘broad discretion’ as to whom to prosecute.” Id. at 607. The majority opinion incorporated into this Fifth Amendment case “ordinary equal protection standards” under the Fourteenth Amendment, id. at 608, requiring the petitioner to show that the government’s enforcement system “had a discriminatory effect and that it was motivated by a discriminatory purpose.” Id. at 608–09 (citing Feeney, 442 U.S. 256; Arlington Heights v. Metro. Hous. Dev. Corp., 429 US. 252 (1977); Davis, ). Justice Powell rejected the petitioner’s claim by emphasizing that they would not only have to show that the government’s enforcement policy had a foreseeable adverse effect but that it had a discriminatory purpose within the meaning of the Feeney standard. See id. at 610 (referencing Feeney’s definition of discriminatory purpose).

Because discretion is essential to the criminal justice process, we would demand exceptionally clear proof before we would infer that the discretion has been abused . . . . Accordingly, we hold that the Baldus study is clearly insufficient to support an inference that any of the decisionmakers in McCleskey’s case acted with discriminatory purpose.73

III. HOSTILITY TO STATISTICAL PROOF OF DISCRIMINATION AND REASONING ABOUT RACE IN MCCLESKEY

72 McCleskey v. Kemp, , 297 (1987); see also id. at 313. 73 Id. at 297. 74 Id. at 313 (“Where the discretion that is fundamental to our criminal process is involved, we decline to assume that what is unexplained is invidious. In light of the safeguards designed to minimize racial bias in the process, the fundamental value of jury trial in our criminal justice system, and the benefits that discretion provides to criminal defendants, we hold that the Baldus study does not demonstrate a constitutionally significant risk of racial bias affecting the Georgia capital sentencing process.”).

75 Id. at 315–17.

76 Id. at 318–19. 77 Id. at 339 (Brennan, J., dissenting). 78 839 F.2d 302, 310 (7th Cir. 1988) (rejecting a pattern-and-practice challenge to Sears’ hiring and promotion practices of female employees where “[v]irtually all the proof offered by the EEOC . . . [was] statistical in nature, or related to the statistical evidence” (alteration in original) (internal quotation marks omitted) (quoting EEOC v. Sears, Roebuck & Co., 628 F. Supp. 1264, 1285 (N.D. Ill. 1986 .

79 490 U.S. 642, 650 (1989) (reversing a finding of prima facie disparate impact where “the Court of Appeals relied solely on respondents’ statistics showing a high percentage of nonwhite workers in the cannery jobs and a low percentage of such workers in the noncannery positions”).

80 488 U.S. 469, 499 (1989) (striking down Richmond’s remedial plan requiring that prime contractors subcontract with a set percentage of minority business enterprises partially because it might “give local governments license to create a patchwork of racial preferences based on statistical generalizations about any particular field of endeavor”).

81 See Tamar Lewin, Statistics Have Become Suspect in Sex Discrimination Cases, N.Y. TIMES (Feb. 9, 1986), http://www.nytimes.com/1986/02/09/weekinreview/statistics-have-become-suspect-in-sexdiscrimination-cases.html [https://perma.cc/J96P-7LS9]. Lewin writes:

Because groups—black, white, Hispanic, male, and female—do not necessarily have the same distribution of, among other characteristics, skills, interest, motivation, and age, a fair shake system may not produce proportional representation across occupations and professions, and certainly not at any

The E.E.O.C.’s charges were based entirely on statistical evidence that women were not hired or promoted into commission sales jobs as frequently as men. The commission did not present any witnesses to testify that they personally had been discriminated against. Although the use of statistics to show discrimination is standard in large class actions - usually supplemented by testimony from individuals - the approach is viewed with disfavor by the Reagan Administration. Indeed, Clarence Thomas, the head of the commission, has said he does not believe in statistical cases.

And while the commission’s lawyers stayed with the Sears case even after their boss disavowed it, civil rights advocates do not expect any more such suits during the Reagan years. “This Administration rejects the well-established practice of using statistics to prove discrimination,” said Isabelle Katz Pinzler, director of the American Civil Liberties Union’s Women’s Rights Project. “They want the only cases to be ones where there is a smoking gun, where someone said, ‘I won’t promote you because you’re a woman or a black.’ Since the days are long past when such cases existed, this approach is effectively a way of repealing the equal employment laws.”

Id.

82 See, e.g., HERITAGE FOUND., Ending Discrimination in Civil Rights, in A MANDATE FOR LEADERSHIP REPORT: AGENDA ’83, at 206, 208 (Richard N. Holwill ed., 1983) (criticizing the continued use of the “effects test” and calling for a new evidentiary standard in which “an uneven result is not necessarily an inequitable result”).

83 McCleskey v. Kemp, , 287 (1987). 84 OFFICE OF LEGAL POLICY, U.S. DEP’T OF JUSTICE, REDEFINING DISCRIMINATION: “DISPARATE

IMPACT” AND THE INSTITUTIONALIZATION OF AFFIRMATIVE ACTION, at i (1987).

given time. This uneven distribution, however, is not necessarily the result of discrimination.85

85 Morris B. Abram, Affirmative Action: Fair Shakers and Social Engineers, 99 HARV. L. REV. 1312,

1315 (1986).

86 Reva B. Siegel, The Racial Rhetorics of Colorblind Constitutionalism: The Case of Hopwood v. Texas, in RACE AND REPRESENTATION: AFFIRMATIVE ACTION 29, 58–61 (Robert Post & Michael Rogin eds., 1998).

87 Brief Amicus Curiae of the Washington Legal Foundation and the Allied Educational Foundation in Support of Respondent at 13–14, McCleskey, (No. 84-6811) (“Petitioner’s theory holds that any deviation from statistically-based norms of racially proportional outcomes in a capital sentencing system would ‘require the invalidation of that system as a whole.’”).

88 Id. at 16. 89 Id. at 17. 90 Id. at 11. 91 McCleskey v. Kemp, , 314–19 (1987).

92 Id. at 319. 93 Id. at 339 (Brennan, J., dissenting). 94 428 U.S. 153 (1976). 95 McCleskey, 481 U.S. at 319 (alteration in original) (quoting Gregg, 428 U.S. at 199 n.50). Gregg invoked realism in rejecting a claim under Furman v. Georgia, 408 U.S. 238 (1972), that “opportunities for discretionary action” in Georgia’s capital punishment scheme violated the Eighth Amendment. See Gregg, 428 U.S. at 199. In Gregg, the Court explained that: “[t]he petitioner’s argument is nothing more than a veiled contention that Furman indirectly outlawed capital punishment by placing totally unrealistic conditions on its use.” Id. at 199 n.50. Thus, Gregg offers no general principle of the kind Justice Powell seems to impute to it.

96 A belief that social stratification is more often explained by group differences rather than by discrimination fueled conservative critique of antidiscrimination law. See supra notes 82–85 and accompanying text. This is the tacit argument of the Washington Legal Foundation brief discussed supra notes 87–90 and accompanying text. Justice Powell seemed to have these considerations in mind. While working on early drafts of his opinion, Justice Powell sent one of his law clerks to gather Federal Bureau of Investigation (FBI) statistics showing that most murders were committed by black people against black victims. In a November 8, 1986 memo to the clerk, Justice Powell wrote:

If and when you have an hour or two, take a look at the FBI’s report on “Crime in the United States” to see whether any of the statistical studies or charts may be of interest. Studies we initiated when I was on the Crime Commission indicated that most murders—at least percentage wise and I think also in absolute terms—were committed by blacks on blacks. Also, in general, most of the crimes committed by blacks were on black victims. The FBI study should also show the percentage of major crimes committed by blacks in comparison with the number of such crimes committed by whites, and with respect to both races in comparison with the state or metropolitan area population figures.

School

Memorandum from Lewis F. Powell, Jr., Assoc. Justice, Supreme Court of the U.S., to Leslie 2–3 (Nov. 8, 1986) (located in Justice Powell’s McCleskey v. Kemp Case File on file with Washington & Lee University 133–35), Law https://scholarlycommons.law.wlu.edu/cgi/viewcontent.cgi?article=1168& context=casefiles [https://perma.cc/CJ2D-JHD9]. The requested copies of the FBI reports can also be found in Justice Powell’s files, with the relevant sentences underlined. See, e.g., FED. BUREAU OF INVESTIGATION, U.S. DEP’T OF JUSTICE, UNIFORM CRIME REPORTS: CRIME IN THE UNITED STATES (July 27, 1986) (located in Justice Powell’s McCleskey v. Kemp Case File on file with Washington & Lee

McCleskey’s arguments are best presented to the legislative bodies. It is not the responsibility—or indeed even the right—of this Court to determine the appropriate punishment for particular crimes. It is the legislatures, the elected representatives of the people, that are “constituted to respond to the will and consequently the moral values of the people.” Legislatures also are better qualified to weigh and “evaluate the results of statistical studies in terms of their own local conditions and with a flexibility of approach that is not available to the courts.”101

University https://scholarlycommons.law.wlu.edu/cgi/viewcontent.cgi? article=1168&context=casefiles [https://perma.cc/CJ2D-JHD9].

School

Law

211–12),

97 See supra note 95 and accompanying text. 98 McCleskey, 481 U.S. at 319 (internal quotation marks omitted) (quoting Furman v. Georgia,

408 U.S. 238, 383 (1972) (Burger, C.J., dissenting .

99 , 248 (1976). 100 443 U.S. 449, 489 (1979) (Powell, J., dissenting). 101 McCleskey, 481 U.S. at 319 (first quoting Furman, 408 U.S. at 383 (Burger, C.J., dissenting); then quoting Gregg v. Georgia, 428 US. 153, 186 (1976 ; see Siegel, Equality Divided, supra note 4, at 20–23 (discussing how the Burger Court developed discriminatory purpose doctrine as a doctrine of judicial deference to political ordering); id. at 23 (“In the 1970s, conservatives who sought to limit the role of federal courts in enforcing the Equal Protection Clause advocated deference to representative government, urging that plaintiffs redirect their [race discrimination] claims to the legislative arena.”)

102 See, e.g., ELIZABETH HINTON, FROM THE WAR ON POVERTY TO THE WAR ON CRIME: THE MAKING OF MASS INCARCERATION IN AMERICA 308–09 (2016) (discussing how “the Reagan administration built upon the strategies its predecessors pursued for the War on Crime, including the militarization of local police forces, the criminalization of social programs, and mass incarceration”).

103 The McCleskey decision occurred at a time of unprecedented growth for the American criminal justice system. During the mid-1980s, “over 90% of all the states in the United States were at the high point of the century for rates of imprisonment,” Franklin E. Zimring, The Scale of Imprisonment in the

United States: Twentieth Century Patterns and Twenty-First Century Prospects, 100 J. CRIM. L. & CRIMINOLOGY 1225, 1232 (2010), and over the course of the decade “the incarcerated population more than doubled in size across all three levels [federal, state, and jail],” NAT’L RESEARCH COUNCIL, THE GROWTH OF INCARCERATION IN THE UNITED STATES: EXPLORING CAUSES AND CONSEQUENCES 39 (Jeremy Travis et al. eds., 2014). This growth was particularly pronounced when viewed in comparative international perspective. See JESSICA JACOBSON ET AL., PRISON: EVIDENCE OF ITS USE AND OVER-USE FROM AROUND THE WORLD 4 (2017) (demonstrating the marked increase in the U.S. prison population compared to ten other nations). As McCleskey made its way to the Supreme Court, America’s booming prison population was a focus of national attention. See, e.g., Arthur S. Brisbane, Crime and Punishment: The Push for Prisons, WASH. POST, Mar. 3, 1985, at A01, A14 (describing the “unmanageable horde” of inmates as the national “prison population soared”); Editorial, The Prison Population, WASH. POST, Aug. 17, 1985, at A18 (noting that America’s prisons had a “population explosion in the last dozen years”); Saundra Saperstein, District Not Alone in Prison Crisis: 35 States Under Court Orders to End Inmate Overcrowding, WASH. POST, July 20, 1986, at A1 (observing that “the number of inmates in the nation is at an all-time high— more than . . . double the figure a decade before”).

104 In the years before McCleskey, commentators highlighted the severe racial disparities apparent in the growing numbers of the incarcerated. See, e.g., Black Men More Likely to Serve Time in Prison, N.Y. TIMES: AROUND THE NATION, July 29, 1985 (highlighting that “[b]lack men are six times more likely than white men to serve in a state prison,” according to a Justice Department report); Bob Keeler, Study: Prison Population to Grow, NEWSDAY, Aug. 27, 1985, at 17 (discussing the overrepresentation of minorities in New York prisons and sentencing laws that are “particularly harsh on minorities”); Courtland Milloy, A Black Criminal Profile, WASH. POST, July 22, 1986, at B3 (discussing prison reform and how “the factor of race must be dealt with head-on . . . [because] blacks in America, while only oneeighth of the population, are convicted of half of all rapes, robberies, and murders”); Wendell Rawls, Jr., Crises and Cutbacks Stir Fresh Concerns on Nation’s Prisons, N.Y. TIMES, Jan. 5, 1982 (highlighting that in addition to higher imprisonment rates, black individuals face “harsher sentences than others who have committed similar crimes”).

105 See Haney-López, supra note 66, at 1783 (“Theoretically, the existence of an illicit purpose is the touchstone of the doctrine. In practice, however, the requirement that malice be proved is so exacting that, since this test was announced in 1979 [in Feeney], it has never been met—not even once.”); Siegel,

Why Equal Protection No Longer Protects, supra note 4, at 1139–41 (illustrating how “courts now use Feeney’s definition of discriminatory purpose to justify a decision to uphold facially neutral state action that has a disparate impact on protected classes”).

106 959 F. Supp. 2d 540 (S.D.N.Y. 2013). In addition to statistical evidence, the plaintiffs submitted evidence that the highest-ranking New York Police Department (NYPD) officer identified “young black and Hispanic youths [aged] 14 to 20” as the targeted population for stop and frisk. Id. at 603.

107 Randall L. Kennedy, McCleskey v. Kemp: Race, Capital Punishment, and the Supreme Court, 101 HARV. L. REV. 1388, 1420–21 (1988) (concluding that “it is unlikely that the doctrine of purposeful discrimination will ever facilitate judicial intervention against the sort of conduct delineated by the Baldus study”).

108 Henry Louis Gates, Statistical Stigmata, 11 CARDOZO L. REV. 1275, 1282 (1990) (writing that McCleskey makes it “doubtful whether statistical evidence—no matter how strong—can ever consitute [sic] clear and compelling evidence of purposeful discrimination”).

109 Aziz Z. Huq, The Consequences of Disparate Policing: Evaluating Stop and Frisk as a Modality of Urban Policing, 101 MINN. L. REV. 2397, 2453–55 (2017) (stating that “McCleskey established a nearinsurmountable barrier” to proving an equal protection violation through statistics of disparate impact and “[a]bsent the miraculous happenstance of testimonial or documentary evidence of bias—a stroke of luck that befell plaintiffs in the challenge to New York’s SQF policy [Floyd]— McCleskey means that the courthouse door is effectively shut to discriminatory-purpose challenges in the criminal justice context”). 110 Selmi, supra note 8, at 213 n.74, 215 (describing McCleskey as the “the most vivid illustration” of a hostility to statistical evidence that renders “further progress challenging statistical disparities through litigation . . . unlikely”).

111 See, e.g., Sharad Goel, Maya Perelman, Ravi Shroff & David Alan Sklansky, Combatting Police Discrimination in the Age of Big Data, 20 NEW CRIM. L. REV. 181, 200, 202 (2016) (“With few exceptions, lower court decisions addressing statistical proof of police discrimination echo McCleskey’s reluctance to use statistical evidence to infer discriminatory intent. . . . The bottom line is that unless officers admit to racial animus, or perhaps if the disparate impact is as striking as in Yick Wo, an equal protection challenge is likely to fail for lack of proof that the police in this particular case or set of cases acted with discriminatory purpose.”); see also John M. Powers, State v. Robinson and the Racial Justice Act: Statistical Evidence of Racial Discrimination in Capital Proceedings, 29 HARV. J. RACIAL & ETHNIC JUST. 117, 128 (2013) (“[McCleskey] eliminated the possibility that defendants could make a cognizable claim under either the Eighth or Fourteenth Amendments based on statistical evidence of racial discrimination in charging or sentencing decisions. McCleskey’s burden of proof . . . is generally acknowledged to be impossible to meet.”).

112 See supra note 7 and accompanying text; Bagenstos, supra note 8 (manuscript at 9) (“Indeed, at a moment in history when overt racism—seen in the reaction among some to the election of a black president, and in a significant part of the movement that elected Donald Trump—once again seems a

major factor in our public life, the suggestion that implicit bias is the central problem may be particularly misleading.”).

113 In his opinion, Justice Powell emphasized that McCleskey “relie[d] solely on the Baldus study.” McCleskey v. Kemp, , 293 (1987) (emphasis added). And in the early stages of drafting the opinion, Justice Powell also wrote in his notes that “Criminal cases should not be decided on basis of statistics alone.” Note of Lewis F. Powell, Jr., Assoc. Justice, Supreme Court of the U.S. 4 (Oct. 1986) (located in Justice Powell’s McCleskey v. Kemp Case File on file with Washington & Lee University School 103–07), https://scholarlycommons.law.wlu.edu/cgi/viewcontent.cgi?article=1168&context=casefiles [https://perma.cc/95EF-CNQZ].

Law

Perhaps noting this language, some courts have suggested a willingness to consider statistical evidence of discriminatory purpose as long as it is coupled with circumstantial evidence. See, e.g., Chavez v. Illinois State Police, 251 F.3d 612, 648 (7th Cir. 2001) (“In this context, statistics may not be the sole proof of a constitutional violation and neither Chavez nor Lee have presented sufficient non-statistical evidence to demonstrate discriminatory intent.” (emphasis added . In addition, some courts have rejected particular statistics offered by litigants but signaled that they might be willing to consider more persuasive or well-constructed statistical analyses. See, e.g., United States v. Laneham, No. CV-16-2930 JB, 2017 WL 4857437, at *27 (D.N.M. Oct. 25, 2017); United States v. Johnson, 122 F. Supp. 3d 272, 360 (M.D.N.C. 2015).

114 Whereas the Baldus study in McCleskey applied to death sentences across the state of Georgia, some courts have expressed a willingness to consider more targeted statistics that apply to the specific prosecutorial officer or entity at issue in the case—such as a county prosecutor’s office, individual prosecutor, or police officer. See, e.g., Belmontes v. Brown, 414 F.3d 1094, 1127 (9th Cir. 2005) (“Unlike McCleskey, . . . Belmontes offered statistics that provided information limited to the charging entity . . . .”); Jefferson v. Terry, 490 F. Supp. 2d 1261, 1340 (N.D. Ga. 2007) (“[U]nlike in McCleskey, Petitioner presents evidence pertaining to the specific decision maker in his case.”); Commonwealth v. Lora, No. 20020413, 2003 WL 22350945, at *13 (Mass. Super. Ct. Sept. 12, 2003) (“[C]ourts are more apt to accept statistics as evidence of invidious discrimination when . . . the sample focuses on decisions made by individuals rather than . . . by departments or branches . . . .”). But in two of those cases, the court still found that there was no equal protection violation because the prosecutor could offer a raceneutral, legitimate motive. See Belmontes, 414 F.3d at 1128–29; Jefferson, 490 F. Supp. 2d at 1341.

115 See, e.g., Mehta v. Village of Bolingbrook, 196 F. Supp. 3d 855, 864 (N.D. Ill. 2016) (“A jury could conclude, based on these statistics, that Bolingbrook police conduct their law enforcement duties differently when confronted with members of racial, ethnic, and religious minorities.”); Smith v. City of Chicago, 143 F. Supp. 3d 741, 754–56 (N.D. Ill. 2015) (referring to a series of statistics and noting that “[p]laintiffs have more than adequately alleged a plausible equal protection claim in relation to the

CONCLUSION: SECURING EQUAL TREATMENT, WITH OR WITHOUT COURTS

purportedly unconstitutional stop and frisk policy”); Floyd v. City of New York, 959 F. Supp. 2d 540, 562 (S.D.N.Y. 2013) (“Both statistical and anecdotal evidence showed that minorities are indeed treated differently than whites.”).

116 See, e.g., Reva B. Siegel, Community in Conflict: Same-Sex Marriage and Backlash, 64 UCLA L. REV. 1728 (2017) (describing how courts can engender new understandings of equality rights even under conditions of backlash).

117 478 U.S. 186 (1986). 118 Cf. Robert C. Post & Reva B. Siegel, Legislative Constitutionalism and Section Five Power: Policentric Interpretation of the Family and Medical Leave Act, 112 YALE L.J. 1943, 1947, 2026–32 (2003) (observing that the Constitution has “multiple interpreters, both political and legal” and discussing the distinctive goods that courts and representative government each contribute to enforcing the Constitution).

119 Numerous scholars, led by Linda Krieger, have drawn on social science evidence of implicit bias to propose new liability standards for federal employment discrimination law. See, e.g., Linda Hamilton Krieger, The Content of Our Categories: A Cognitive Bias Approach to Discrimination and Equal Employment Opportunity, 47 STAN L. REV. 1161, 1211 (1995) (“The assumptions underlying Title VII’s disparate treatment theory have been so substantially undermined by social cognition theory that they can no longer be considered valid.”). It appears that in twenty years two Title VII cases have endorsed the notion of implicit bias. See Thomas v. Eastman Kodak Co., 183 F.3d 38, 42 (1st Cir. 1999); Kimble v. Wisconsin Dep’t of Workforce Dev., 690 F. Supp. 2d 765, 775–76 (E.D. Wis. 2010). Kimble has not been followed since it was decided in 2010.

For examples of commentators applying implicit bias work in the constitutional context, see Ivan E. Bodensteiner, The Implications of Psychological Research Related to Unconscious Discrimination and Implicit Bias in Proving Intentional Discrimination, 73 MO. L. REV. 83, 127 (2008) (calling for modifications to burdens of proof) and Reshma M. Saujani, “The Implicit Association Test”: A Measure of Unconscious Racism in Legislative Decision-Making, 8 MICH. J. RACE & L. 395, 422 (2003) (“Plaintiffs should be able to use the Implicit Association Test as circumstantial evidence within the Arlington Heights factors to demonstrate that the challenged practice was not constitutionally justified and relief is warranted because of unconscious prejudice on the part of the decision-maker.”), and compare Eva Paterson et al., The Id, the Ego, and Equal Protection in the 21st Century: Building upon Charles Lawrence’s Vision to Mount a Contemporary Challenge to the Intent Doctrine, 40 CONN. L. REV. 1175, 1191 (2008) (“While the ultimate goal of dismantling the Intent Doctrine may, at present, appear somewhat remote, legal scholars and practitioners have made significant headway in recent years towards chipping away at its foundations. In several areas of law, litigators have begun marshalling social science evidence around implicit and institutional bias and using this evidence to help prove instances of discrimination.”).

120 See Jerry Kang, The Missing Quadrants of Antidiscrimination: Going Beyond the “Prejudice Polygraph”, 68 J. SOC. ISSUES 314 (2012). Jerry Kang urges use of implicit bias research, not only to prove discrimination but to stimulate self-awareness among a range of decisionmakers with authority and influence. Id. at 315.

121 For example, social psychologist Lorie Fridell has become a leader in addressing biased policing. She conducts trainings at law enforcement agencies across the country, aiming to use the science of bias to educate officers about the effects of unconscious bias and to teach them about managing and reducing their biases. See Lorie Fridell, This Is Not Your Grandparents’ Prejudice: The Implications of the Modern Science of Bias for Police Training, TRANSLATIONAL CRIMINOLOGY, Fall 2013, at 10, 11 (2013). Professor Phillip Atiba Goff, an expert in contemporary forms of racial bias and discrimination, also conducts interventions involving implicit bias training in police departments across the nation. Megan Harris, Meet the Man Helping Pittsburgh Police Confront Their Racial Biases, PITTSBURGH NPR (Sept. http://wesa.fm/post/meet-man-helping-pittsburgh-police-confront-their-racial-biases 21, [https://perma.cc/94RT-KAJC]. In addition, in a study that received widespread attention, neuroscience and psychology professor Josh Correll developed a simulator to study and explore the impacts of police officers’ implicit biases in decisions to shoot. See Tom James, Can Cops Unlearn Their Unconscious Biases?, ATLANTIC (Dec. 23, 2017), https://www.theatlantic.com/politics/archive/2017/12/implicit-biastraining-salt-lake/548996 [https://perma.cc/A9AF-9YJP]. The Justice Collaboratory at Yale Law School combines procedural justice and implicit bias training for police officers. See, e.g., Megan Quattlebaum, Presentation to the Connecticut Police Training Task Force, CONN. GEN. ASSEMBLY (Dec. 9, 2016),

2017),

https://www.cga.ct.gov/ps/tfs/20161206_Task%20Force%20to%20Study%20the%20Curriculum%20an d%20Education%20of%20Police%20Officers/20161206/Meeting%201%20Powerpoint.pdf [https://perma.cc/7RWJ-PQV9]. Interventions to stem implicit bias are also occurring in sectors outside of policing, including among teachers to combat school-to-prison pipelines and among judges and attorneys to help debias perceptions and improve interactions. See Implicit Bias Toolbox, AMERICAN BAR ASSOCIATION, https://www.americanbar.org/groups/litigation/initiatives/task-force-implicitbias/implicit-bias-toolbox.html for . . . exploring implicit bias and approaches to ‘debiasing’” aimed at attorneys and judges); Juvenile Justice and Education, CENTER FOR POLICING EQUITY, http://policingequity.org/juvenile-justice-education [https://perma.cc/PKZ5-YS2D] (describing the Center for Policing Equity’s work in developing an implicit bias training for educators and school resource officers to combat school-to-prison pipelines).

[https://perma.cc/9QYN-CW7P]

(describing a “Toolbox

122 See, e.g., U.S. DEPT. OF JUSTICE, THE CIVIL RIGHTS DIVISION’S PATTERN AND PRACTICE POLICE REFORM WORK: 1994–PRESENT (2017) (describing how the Civil Rights Division’s police reform agreements with the Baltimore Police Department, Newark Police Department, Ferguson Police Department, Cleveland Police Department, Los Angeles County Sheriff’s Department, New Orleans Police Department, and East Haven Police Department all require officers to undergo implicit bias training, as it is “a core feature of the Division’s reform agreements”); Press Release, U.S. Dept. of Justice, Department of Justice Announces New Department-Wide Implicit Bias Training for Personnel (June 27, 2016), https://www.justice.gov/opa/pr/department-justice-announces-new-department-wideimplicit-bias-training-personnel [https://perma.cc/RKP4-FAB3] (announcing that the Department of Justice made implicit bias training mandatory for its own law enforcement agents and prosecutors); J. David Goodman & Al Baker, De Blasio Kept Crime Down in First Term. His Next Goal: Nicer Police, N.Y. TIMES (Nov. 5, 2017), https://www.nytimes.com/2017/11/05/nyregion/bill-deblasio-policecrime.html [https://perma.cc/UD4C-J9UZ] (noting New York City Mayor Bill de Blasio’s plan to add implicit bias training to the NYPD training curriculum); Nicole D. Porter, Racial Impact Statements, SENTENCING PROJECT (Dec. 1, 2014), https://www.sentencingproject.org/publications/racial-impactstatements [https://perma.cc/NEM6-8LM5] (describing how a few states have begun evaluating the potential disparities of proposed legislation prior to adoption and implementation by using racial impact statements).