Courts usually don’t grant habeas claims for criminal defendants who allege that their lawyer’s racism prejudiced their defense unless the racial animus is obvious on the cold trial record. In Ellis v. Harrison, the Ninth Circuit had the opportunity to relax this standard and grant habeas relief to a client of a known racist lawyer, Donald Ames, even though evidence of Ames’s racism was not apparent on the trial record. While the court did grant habeas relief, it did so without much precedential effect, leaving the remaining clients of Donald Ames and other racist lawyers, whose animus remains unknown to the court, without a remedy. While ten of the eleven en banc judges in Ellis voted to grant habeas relief, only four joined a concurrence that would have created a per se rule of prejudice when a lawyer is found to harbor extreme racial animus towards their client’s racial group despite both the State and the defendant’s arguments in favor of the per se rule. Why?

I argue that, contrary to popular theories of judicial decision-making, courts facing hard cases where equitable and doctrinal factors weigh in opposite directions turn to procedural mechanisms to arrive at equitable outcomes that do not create binding doctrinal precedent. Rather than reaching the merits of the claim, the Ellis en banc court granted habeas because the State and the defendant agreed that habeas should be granted. Despite granting habeas, the court has created conflicting criteria for granting future relief that at once depend on the State’s agreement with, and opposition to, granting habeas, effectively blocking many future claims. In light of the procedural precedent that Ellis created for granting habeas where the State and the defendant agree, I argue that attorneys general in the Ninth Circuit should proactively review other cases involving racist lawyers and support habeas relief where the judiciary has thus far refused to do so.

Introduction

Should our legal system assume that a lawyer who refers to his Black clients as “n[*]rs” and an Asian-American judge as a “f[]king J[*]p” who should “remember Pearl Harbor” can competently represent clients of color? Donald Ames said all those things and more while representing Ezzard Ellis and other Black clients as a public defender in high stakes criminal trials, including death penalty cases.[2] Should Ellis and Ames’s other clients of color be entitled to new trials even if Ames’s racial animus towards them was not apparent on the cold trial record?

Ellis v. Harrison presented a rare opportunity for the Ninth Circuit to act as “occasional legislators”[3] and create a rule of per se prejudice where a court finds that a client’s lawyer harbors “deeply held discriminatory animus against the defendant’s racial group.”[4] While a previous three-judge Ninth Circuit concurrence invited the creation of a rule granting new trials for clients of virulently racist lawyers, the en banc court balked at implementing a rule with such wide-reaching implications on grounds ranging from the procedural posture of the litigants to the health of the court system as a whole.

Ultimately, the en banc court issued a two paragraph majority decision that granted Ellis habeas relief on the grounds that the State and the defendant agreed on the outcome and directed the district court to grant a conditional writ of habeas corpus.[5] The court justified its summary grant of habeas relief by pointing to California’s decision to support habeas relief just before the en banc hearing after arguing for the denial of habeas relief for years prior.[6] After this about-face, California agreed with Ellis that the court should grant habeas because Ames’s racial animus towards Black Americans denied Ellis his right to counsel under the Sixth Amendment.

Despite the dissent’s implication that Ellis stands for a “new constitutional rule,”[7] the en banc oral argument and the court’s resort to equitable relief reveals the limit of the court’s ability to provide relief for clients of racist lawyers. Judge Jacqueline Nguyen’s concurrence, which explained the legal conclusions the majority must have adopted to grant relief, reveals how narrow the court’s holding truly is.[8] In fact, it is unlikely that even Ames’s other Black clients will be able to take advantage of the legal framework articulated by Nguyen’s concurrence.

Contrary to popular theories of judicial decision-making, the record in Ellis suggests that both substantive and procedural legal norms constrain judicial decision-making even when there is no threat of reversal by a higher court.[9] In this Note, I develop a theory of judicial decision-making to explain how judges exercise discretion in cases where legal doctrine suggests a result that conflicts with intuitive notions of justice. By discretion, I mean what Professor David Shapiro calls normative discretion, defined as “discretion delegated to a rulemaking or adjudicative body [here, the judiciary] by the legislature.”[10] Ultimately, I argue that a judge’s consideration of their policy preferences, which is often considered off-limits when deciding cases, rightfully falls under their power to decide cases in equity.

Following Judge Richard Posner, I conceptualize the range of “acceptable” judicial actions as a two-dimensional field that expands and contracts over the course of a case as judges balance legal doctrine, techniques for applying legal doctrine (e.g., stare decisis), and policy views. In other words, the size of the zone of discretion is formed by the interaction between a judge’s view of the equities in a case and the “existing legal materials” that may compel a given outcome.[11] Further, I argue that while judges balance the vectors of law and equity differently based on their political ideology (and therefore perceive smaller or larger zones of discretion in a given case), legalistic concerns ultimately predominate by limiting the judicial avenues for expressing equitable concerns even where there is no threat of reversal by a higher court. When faced with equitable factors that strongly suggest one result and a legal regime that restricts a court’s discretionary ability to grant that result, courts are likely to fall back on procedural mechanisms as these mechanisms comprise the core of the court’s authority under Article III. I call this “procedural equity” and argue that, while based on the policy preferences of individual judges, such equitable decision-making falls within the authority of Article III courts. If, as my model suggests, relief is unlikely to be granted to other clients of racist lawyers through the courts, attorneys general in the Ninth Circuit should affirmatively support habeas relief or new trials for Donald Ames’s clients and clients of other racist lawyers.

This Note proceeds in three parts. In Part I, I outline the precedents that support the denial of habeas relief for clients of racist lawyers. In Part II, I dissect the Ellis panel’s opinions and the en banc opinions. Finally, in Part III, I review existing scholarship on judicial decision-making before articulating a synthesis of two existing theories—attitudinalism and legalism—and apply it to Ellis.

I. Courts Usually Refrain from Finding that a Lawyer’s Racial Animus Affected Their Clients’ Sixth Amendment rights.

There are three standards used to determine whether a claim of ineffective assistance of counsel rises to a Sixth Amendment violation: Strickland, Cronic, and Sullivan. Proving ineffective assistance of counsel under Strickland requires the plaintiff to show not only that the trial counsel’s performance “fell below an objective standard of reasonableness”[12] but also that “the deficient performance prejudiced the defense.”[13] Cronic and Sullivan describe two mechanisms for satisfying Strickland’s second prong—establishing a linkage between a lawyer’s conduct and deficient performance. Under Cronic, a court can presume prejudice when the defendant is completely denied counsel, satisfying the second Strickland prong without a showing that counsel’s error biased the outcome.[14] Under Sullivan, “an actual conflict of interest [that] adversely affected [the client’s] lawyer’s performance” can satisfy the second Strickland prong.[15]

When racist lawyers are involved, the barrier to habeas relief under Strickland arises from difficulties in establishing that a lawyer’s racism caused the constitutionally deficient performance. Strickland’s second prong only requires the defendant to show “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different,” rather than requiring proof that counsel’s performance more-likely-than-not affected the outcome. [16] The cases I discuss below—first collected by Sheri Lynn Johnson, John H. Blume, and Patrick M. Wilson in Racial Epithets in the Criminal Process—reveal that courts are loath to connect a lawyer’s racist comments to his representation of a particular client unless those comments were made in a courtroom or to the client directly.[17] First, I examine a case where the racist lawyer’s animus was present on the trial record and the court granted habeas relief. Second, I review cases where the racist lawyer’s animus was known to the court but not evident on the cold trial record or expressed to the client, leading courts to deny habeas relief. Lastly, I turn to cases where the racist lawyer’s animus was known to the court but was not present on the trial record, and the court considered habeas relief under either the Cronic or Sullivan standard.

A. Racial Animus Present on the Trial Record

Habeas relief has been granted in cases where a lawyer’s racism manifests in the courtroom during trial proceedings. In State v. Davis, the Supreme Court of Florida granted habeas relief when a racist lawyer with a Black client told potential jurors: “Sometimes I just don’t like [B]lack people. Sometimes [B]lack people make me mad just because they’re [B]lack.”[18] The court found the lawyer’s “expressions of racial animus . . . so seriously affected the fairness and reliability of the proceedings that [its] confidence in the jury’s verdicts of guilt [was] undermined.”[19]

When racial animus is apparent on the trial record, as it was in State v. Davis, courts cannot escape confronting the impact of racial bias on representation. To let stand a verdict where a lawyer used racist language to describe his client in the courtroom casts doubt on the verdict because the lawyer’s animus may infect jurors’ decision-making.

B. Racial Animus Neither Present on the Trial Record nor Directed at the Client

When racial animus is directed at parties outside the courtroom and outside the attorney-client relationship, on the other hand, courts often refuse to grant habeas relief. Courts often turn to procedural mechanisms or rote recitation of the Strickland standard to deny habeas and rarely discuss the epithets at length in majority opinions.

In State v. Yarborough, a racist lawyer told a mitigation expert that “he [counsel] knew [B]lack people he liked, but that Kevin [his client] was a n[**]r.”[20] In a limited, if not cursory, inspection of Yarbrough’s claim, the court determined that, while the statement was “grossly inappropriate, offensive, and distasteful,” there was no evidence of ineffective assistance on the trial record and the remark did not affect the outcome of the trial.[21]

In Jones v. Campbell, a racist lawyer told his legal secretary “something to the effect of ‘that n[**]r is going to fry.’”[22] First, the Jones court determined that the allegation of racial bias was procedurally defaulted because Jones only included it in the introduction and conclusion of his state and federal habeas petitions.[23] Second, the court found that Jones’s claim was without merit regardless of the procedural default because Jones did not provide any evidence that his lawyer made racial comments to Jones or during his representation of Jones.[24]

Yarborough and Jones emphasize the importance of finding a direct link between racial epithets and the client in a specific case to the success of habeas petitions under Strickland. Courts’ refusal to consider the pernicious, invisible effects of racism on representation may be a reaction to the difficulty of constraining a rule that would grant habeas relief solely on the basis of a lawyer’s racist actions outside the courtroom absent any clear representational failure on the cold record or indicia of the defendant’s factual innocence.

C. Habeas Relief Under Cronic and Sullivan

In some instances, courts presume prejudice under the second prong of Strickland when counsel is denied or labors under a conflict of interest.

Under Cronic, the second prong of Strickland is met when counsel is completely denied; the defendant need not show that his counsel’s error biased the outcome.[25] Complete denial of counsel can occur when the counsel is absent for the proceedings or “fails to subject the prosecution’s case to meaningful adversarial testing.”[26] In Frazer v. United States, the Ninth Circuit granted habeas to a client whose racist lawyer hurled racial invectives at him. The court theorized that the lawyer broke his duty of loyalty to his client and “a presumption of prejudice [would be] appropriate without inquiry into the actual conduct of the trial.”[27] The racist lawyer called his client a “stupid n[**]r son of a bitch” and told his client he would “find him to be very ineffective” if he went to trial.[28] The Frazer court determined that these statements would render all legal advice given thereafter “fatally suspect,” thereby causing a “total lack of communication” resulting in ineffective assistance of counsel.[29]

Under Sullivan, “actual” conflicts of interest “adversely” affecting an attorney’s performance may also meet the second prong of Strickland.[30] For example, a conflict of interest could occur when the same defense lawyer represents two defendants whose innocence depend on conflicting interpretations of the facts. In Mickens v. Taylor, the Supreme Court reemphasized that a conflict did not create a Sixth Amendment violation unless the defendant could establish that the conflict actually affected the adequacy of representation.[31]

In Mayfield v. Woodford, however, another of Ames’s clients argued that Ames’s racism constituted a conflict of interest that prevented Ames from representing him.[32] An en banc Ninth Circuit panel found that Ames had rendered ineffective assistance at the penalty phase of Mayfield’s trial and granted habeas relief on that basis. Although Mayfield informed the en banc court of “six declarations indicating that Ames was racially prejudiced,” two of which “related racial epithets that Ames used in reference to minority clients,”[33] the court refused to grant habeas relief on the basis of Ames’s racial prejudice.[34] Specifically, Ames’s former legal secretary alleged that Ames “‘consistently’ referred to his [Black] clients as ‘n[]rs’ . . . and called a fellow lawyer ‘a big [B]lack n[]r trying to be a white man.’”[35] Another former employee of Ames testified that “because his client . . . was [B]lack [Ames] did not trust him and did not care what happened to him.”[36] Ultimately, the court denied Mayfield a Certificate of Appealability based on a conflict of interest arising from Ames’s racial prejudice. The denial was both because the court found that Sullivan did not accommodate conflicts premised on racial bias and because Mayfield had not demonstrated a causal link between Ames’s racism and his poor performance as counsel.[37] The court minimized the pattern of racist statements by pointing out that none of them were used against Mayfield himself and emphasized that Mayfield “did not demonstrate that [Ames] performed deficiently because of his racial bias.”[38]

In a vigorous dissent, Judge Susan Graber concluded that Ames “actively served the interests of the prosecution” by introducing evidence and making arguments that aided the prosecution’s case, including during closing argument.[39] She concluded that “this is one of the rare cases in which ‘defense’ counsel’s sympathies so obviously and cynically belonged to the prosecution that Petitioner received the equivalent of no counsel at all”[40] despite a lack of obvious racial animus on the trial record.

In the end, the discussion of racial animus was beside the point in Mayfield. Mayfield was granted habeas relief on the basis that he was ineffectively represented during sentencing, but the Mayfield majority refused to consider whether Ames’s racism affected his representation absent a racist statement made by Ames in the courtroom or directly to Mayfield.[41] Unlike Ellis, Mayfield created precedent that found Ames’s racist statements too attenuated to grant habeas relief—precedent that the Ellis majority did not formally overturn. But, like Ellis, Mayfield suggests that courts turn to procedural, equitable mechanisms for providing habeas relief to avoid habeas relief based on a lawyer’s racial animus. Courts turn to procedural, equitable mechanisms for providing habeas relief to avoid granting relief based on a lawyer’s racial animus.[42]

II. The Ellis Court Pushed Doctrine Further on Procedural Grounds by Granting Habeas Without Evidence of Racial Animus on the Cold Trial Record

Ellis v. Harrison is a quintessential “hard case” involving a racist lawyer. Ames’s racism was not obvious on the trial record, and Ellis did not raise a claim predicated on Ames’s racism until years after his conviction when he learned of Ames’s “frequent use of deprecating remarks and racial slurs about his clients” in a newspaper article.[43] When other courts have faced overturning cases involving a racist defense lawyer but without an epithet directed at the lawyer’s client, they often find that the lawyer’s racism is too attenuated to prejudice the outcome of the case and deny habeas relief because Strickland’s second prong is not satisfied.[44] Following Mayfield, a unanimous three-judge panel in Ellis initially denied habeas on precisely this ground.[45] Judge Nguyen authored a separate concurring opinion that all three judges joined. The opinion found that Ames’s racism could have affected his performance in insidious ways that were invisible on the cold record but still constituted effective denial of counsel. Ultimately, however, Judge Nguyen’s concurring opinion determined that the court could not grant habeas relief because the case was indistinguishable from Mayfield.[46]

In Ellis’s en banc decision, the majority granted habeas relief despite the fact that Ames’s racism was not visible on the cold trial record. They did so through a cursory opinion based on the fact that both Ellis and the State advocated for habeas relief before the en banc court.[47] Judge Nguyen again authored a concurrence that rearticulated much of the reasoning from her panel concurrence.[48] Below, I consider Ellis in context and conclude that the habeas relief granted by the en banc panel is unlikely to be repeated in other cases involving clients of other racist lawyers.

In 1991, Ezzard Ellis was convicted of first-degree murder, attempted murder, and robbery after two men were shot and had their car stolen while waiting at a McDonald’s drive-through in San Bernardino County, California.[49] Ellis’s case was tried five times; the first two trials ended in mistrials due to unavailable witnesses, and the second two ended in hung juries.[50] A jury convicted Ellis in his fifth trial and sentenced him to life without parole.[51]

At each trial, Ellis, a Black man, was represented by Ames, a white court-appointed defense attorney.[52] In a sworn declaration, Ames’s daughter recalled a case from around the time of Ellis’s fifth trial involving “African-American men . . . accused of holding up or robbing someone at a fast food restaurant.”[53] Ames’s daughter alleged that Ames used racial slurs to refer to his client and “commented on how stupid his client was,” suggesting that Ames harbored animus towards Ellis in particular.[54] Because of Mayfield, Ames’s racism was well known to the Ninth Circuit before Ellis first learned of it in 2003. That year, Ellis began a lengthy appeals process that culminated in a three-judge Ninth Circuit opinion and subsequent en banc hearing that conditionally granted him habeas relief.[55]

In assessing Ellis’s claim for habeas relief, the three-judge Ninth Circuit panel found that Mayfield compelled a denial of habeas because there, as in Ellis’s case, the petitioner could not show that Ames’s racism caused him to make poor strategic decisions.[56] Interestingly, however, all three judges on the panel signed onto a concurrence authored by Judge Nguyen, which suggested that the court overrule Mayfield.[57] While noting that a “racially tinged comment by defense counsel” should not necessitate overturning a conviction, Nguyen explained that when counsel “as deeply racist as Ames” “expresses such utter contempt and indifference about the fate of his minority clients,” counsel cannot provide “the reasonably competent representation that the Sixth Amendment demands.”[58] Nguyen emphasized that “[t]he deleterious effect of such racism on the outcome is usually impossible to prove and, under these circumstances, we should presume prejudice.”[59] Nguyen ended her concurrence by noting the constraints on the panel’s decision-making authority “[b]ecause I cannot in good faith distinguish Ellis’s case from Mayfield, I reluctantly concur in the opinion. Had we not been bound by Mayfield, I would have granted Ellis’s petition.”[60] The concurrence spurred the court to successfully call Ellis en banc.

Given that the three-judge panel ruled against granting habeas, the success of the en banc call meant that at least half of the active, non-recused judges on the circuit believed that the court should at least reconsider, if not overrule, Mayfield.[61] This suggests that the concurrence’s underlying equitable principle—presuming prejudice when an attorney exhibits extreme racial animus towards a client’s racial group regardless of its visibility on the trial record—carried significant support among at least half of the active circuit court judges.[62] California’s subsequent decision to support habeas suggests that the occurrence convinced the attorney general’s office as well.

A. The Effect of California’s About-Face

Counsel for the State of California supported the denial of habeas relief throughout Ellis’s appeals process and argued for denial before the three-judge panel.[63] After the three-judge panel decision denied habeas and the subsequent successful en banc call, the State took “a fresh look at the legal landscape and at the facts in this case and . . . consider[ed] the panel opinion” before agreeing with Ellis that the court should overrule Mayfield.[64]

At en banc oral argument, both Ellis and the State put forward a new rule that followed the contours of the equitable principle outlined in Nguyen’s panel concurrence.[65] Per the State’s counsel: “When a criminal defense attorney harbors extreme animus towards the defendant’s racial group, it creates a great likelihood that the attorney will perform ineffectively but in ways that are not likely to be apparent on the cold record. Under such circumstances, the court should apply a rule of per se prejudice . . . .”[66] Like Nguyen’s concurrence, both parties emphasized that, to find animus under the rule, a petitioner would need to show an “objective manifestation of such a belief, whether it’s statements of the type that we have here, writings or perhaps documented membership in a hate group”[67] and that an ordinary racist comment would not rise to the level of “extreme animus” that would trigger a presumption of prejudice.[68]

The central puzzle is this: having created a space where the court could act as what Judge Posner refers to as “occasional legislators,”[69] and after both parties invited the Ninth Circuit to implement a rule based on the same principle that garnered enough support for a successful en banc call, nearly every member of the en banc panel balked at implementing the proposed rule during oral argument. Furthermore, in the final en banc Ellis opinion, only three judges signed onto Judge Nguyen’s panel concurrence explaining the court’s legal reasoning. Why?

I argue that the State’s decision to support Ellis’s habeas petition after vigorously opposing it for more than twelve years allowed the court to consider a constitutional expansion of habeas that would account for the pernicious, invisible effects of racism on a lawyer’s performance, but simultaneously created problems of justiciability under Article III. In essence, the State’s decision shrunk the acceptable field of judicial discretion in the eyes of some judges on the en banc panel. In response to these constraints, the court issued a decision based in equity—limiting Ellis’s precedential value.

Although the State’s support of Ellis’s habeas petition removed procedural constraints that would have otherwise prevented the court from hearing the case,[70] it nonetheless introduced new normative legal issues around deciding it. To several members of the court appointed by both Democratic and Republican presidents, the State’s reversal suggested that the case no longer met the “case or controversy” requirement because the parties were no longer truly adverse to one another.[71] Importantly, the fear of reversal could not have influenced the judges’ decision-making; if the court enacted the rule proposed by the State, neither party would have been likely to appeal the decision since both parties would have been happy with the result. Moreover, as Judge Consuelo Callahan suggested during the oral argument, if the court did not follow the parties’ recommendation, it is possible, perhaps likely, that neither party would seek certiorari from the current conservative Supreme Court for fear of a ruling that would prohibit habeas relief in cases like Ellis.[72] Ironically, the absence of the constraint of reversal was cited by judges across the political spectrum as a reason not to implement the suggested rule.[73]

B. Competing Legal Standards and Justiciability Concerns

Ellis’s en banc oral argument and subsequent decision reveal important fault lines among the en banc judges, even those who appeared to agree on the substance of the proposed rule. The majority opinion, signed by ten of the eleven judges, simply granted a conditional writ of habeas corpus “[i]n light of the State’s concession that habeas relief is warranted” and released Ellis from custody unless he was retried “within a reasonable period of time.”[74] Two concurrences attempted to explain the majority’s summary granting of habeas, but in very different ways. Judge Paul Watford’s concurrence—joined by Judges Kim Wardlaw, Andrew Hurwitz, John Owens, and Michael Hawkins—was a response to the dissent’s assertion that 28 U.S.C. § 2254(d) (The Antiterrorism and Effective Death Penalty Act of 1996 or “AEDPA”) prevented the en banc court from considering Ellis’s habeas petition.[75] Judge Nguyen’s concurrence went much further. Joined by Chief Judge Sidney Thomas and Judge Mary Murguia, Judge Nguyen laid out the necessary legal justification underpinning the majority’s grant of conditional habeas. She “strongly disagree[d] with the majority’s refusal to explain its decision, particularly in the face of a vigorous dissent.”[76] Judge Callahan supplied that dissent and argued that Ellis could not demonstrate that Ames’s prejudice had biased his representation under either Strickland or Sullivan, and that the application of Cronic was improper under AEDPA. No other judges joined Callahan’s dissent but, interestingly, the other two judges appointed by a Republican president, Milan Smith and Jay Bybee, joined only the majority opinion and neither of the justificatory concurrences.

In the final decision, as well as the en banc oral argument, the judges fell into essentially two camps. In one camp were those who believed that the rule must include a link between the racist lawyer’s animus and his representation, as required under the second prong of Strickland (hereinafter the Strickland camp). The other camp believed that such a nexus was unnecessary to grant conditional habeas because extreme racial animus constitutes an abdication of counsel’s duty of loyalty to their client under Cronic when the client is of a race his lawyer abhors (hereinafter the Cronic camp).[77] The justifications in the Strickland camp were split. Some judges put forth primarily policy reasons for opposing the per se rule (hereinafter the policy subdivision).[78] Other judges indicated that they believed procedural concerns predominated to the point that discussion of the proposed rule wasn’t properly before the court (hereinafter the justiciability subdivision).[79] This split within the Strickland camp divided policy-driven judges who either dissented or joined no opinion other than the two-paragraph majority from justiciability-concerned judges who joined Judge Watford’s concurrence. The Cronic camp joined Nguyen’s concurrence. What is difficult to explain is why many judges in the Strickland camp—those who believed any proposed rule should require a link between racial animus and representation—still voted in favor of habeas for Ellis when both parties agreed that the nexus did not exist on the trial record.

1. The Strickland Camp

Judges Callahan, Smith, and Bybee—all Bush appointees—fell into the policy subdivision of the Strickland camp. They each emphasized issues with the rule stemming from its assumption that a direct link between racial animus and deficient representation was not required to grant habeas, rather than issues with justiciability.

Judge Callahan, for example, challenged the conceptual basis for assuming that a racist lawyer would not zealously represent a Black client unless there was something evident in the trial record: “[T]he man’s a racist and therefore everything that follows, even though without any link to what happened at the trial . . . or the strength of the evidence, or any injustice that would have occurred, [the defendant] gets to pass.”[80] She further questioned how the court would evaluate a lawyer’s racist conduct if the rule were to go into effect: “Mr. Ames was a bad [person] . . . I’m not sure that if, if that is going to become part of the standard that there would be any counsel out there that you couldn’t find something bad to say about as a human being.”[81]As Judge Nguyen pointed out, however, “we’re not talking about a situation where somebody may have made an error in college by participating blackface at a party or made some stray comments, right? We’re talking about a lawyer who, during the period of representing African-Americans, repeatedly use[d] racial slurs against them, believed in the inherent inferiority of African-Americans, [and] said he didn’t care what happened to them . . . .”[82]

Like Judge Callahan, Judge Smith offered policy arguments attacking the rule and expressed concerns that it could unleash a flood of litigation: “Even now, before we consider any rule here, virtually every capital case we get—the lawyer . . . is accused of committing malpractice.”[83] Judge Smith was especially concerned that the rule did not require any connection between the racial animus and the lawyer’s performance at trial: “[I]t seems like you’re asking for a -- basically a per se rule. So, you don’t have to prove anything. I get your point, especially with respect to Mr. Ames. But you’re really talking about putting a bomb under the whole criminal justice system, to the degree you have a lawyer who has any alleged hint of racism.”[84]

Judge Bybee fell into the policy subdivision as well because of his concern that the Ninth Circuit would be imposing the suggested rule on all courts in the Ninth Circuit, but he fell closest to the justiciability subdivision.[85] Judge Bybee viewed the State’s decision to support Ellis’s petition as a sign that California should impose this rule on itself rather than ask the en banc panel to impose it on every state in the Ninth Circuit: “Why is it more important for the California attorney general to go back to the California courts and ask that the California courts to adopt this rule . . . why isn’t the [S]tate cleaning its own house instead of coming to us?”[86]

Although these judges brought up legal constraints (precedential, in Callahan’s case, and justiciable, in Smith’s) the thrust of their argument was that, in Smith’s words, “[i]t’s just a guessing game”[87] as to whether a lawyer harbors racial animus towards a particular client if there isn’t a direct linkage between the animus and the deficient trial strategy. That is, Judges Callahan, Smith, and Bybee disagreed with the underlying equitable principle in Judge Nguyen’s concurrence that extreme racial animus necessarily leads to deficient representation.[88]

Judges Wardlaw, Hurwitz, Owens, and Watford fell into the justiciability subdivision of the Strickland camp. It is more difficult to tell whether these judges also disagreed with the idea that a nexus between racial animus and deficient representation should be required. They either argued at oral argument that the court could not decide the case because there was no case or controversy or signed both the majority opinion and Judge Watford’s concurrence declaring that the court could grant habeas despite AEDPA, but without giving much legal justification for their vote to grant habeas relief.

Judge Wardlaw (a Clinton appointee) emphasized that the State’s decision not to oppose the petition for habeas meant that she didn’t know “whether this case is even just about—justiciable any longer”[89] because the State would not seek Supreme Court review if the court implemented the proposed rule.[90] Wardlaw was incredulous that the State’s counsel would not admit that the State would not seek certiorari: “I mean, how could you say [‘]we’re not going to oppose it[’]? And [‘]we’re going to argue the same side[’]? And then [the Ninth Circuit panel] come[s] back, we do something that you don’t precisely like, aren’t you now judicially estopped from going before the Supreme Court? . . . You haven’t really thought about that, have you?”[91] Wardlaw explicitly noted that the absence of the reversal constraint “behooves us if we do come down that way to make a narrow rule.”[92]

Of the judges in the Strickland camp, Judge Hurwitz fell closest to the Cronic camp. He stated during oral argument that “I think [the State is] correct on this, our prior case law [meaning Mayfield] is wrong.”[93] He also expressed concern at the breadth of the proposed rule during oral argument. He argued for a narrower rule that would only provide for habeas if the racist lawyer made racist statements about their own client or directed at their own client, rather than racist statements towards any person of the same race as their client.[94] Hurwitz argued “it’s an easy line to draw in this case. Why should we go . . . drawing other lines?”[95] This standard expands Frazer but falls short of the per se rule advocated by the State because it would not cover instances where a racist lawyer makes statements about other people of color but not their specific client.[96]

Despite Judge Hurwitz’s apparent agreement with relaxing the Strickland standard, he spent much of oral argument emphasizing the absence of a case or controversy after the State sided with Ellis: “[Y]ou [the State] have every ability to walk over to the counsel table, extend your hand to counsel and say, ‘We hereby so agree.’ But instead of doing that, you’re saying to us, ‘Even though we agree and have no controversy between us on this, we would like you to adopt a rule that involves some difficult questions.’ . . . [T]hat may be the correct rule . . . [but] [w]hy can’t you guys just go work this out?”[97] Hurwitz went on to note that in another case “in which the [S]tate is less charitable . . . [the court] can confront the issue and decide it” and that the State could grant Ellis equitable relief without any recourse to the state courts by stipulating a judgment against the state.[98]

Wardlaw and Hurwitz’s statements were founded in structural concerns about the role of the federal courts in adjudicating non-adversarial disputes, but these concerns arise only when the two parties agree on an outcome. Conventional wisdom would suggest that the absence of reversal would embolden judges to enact their own preferences,[99] but here the absence of the constraint heightened structural concerns.[100]

Hurwitz’s statements at oral argument coupled with his decision not to sign Nguyen’s en banc concurrence suggest that his equitable concerns in this case were overwhelmed by the legalistic constraints created by the State’s siding with Ellis. First, the parties’ non-adversarial relationship violated Hurwitz’s internal conception of when a judge should decide cases.[101] Second, the State’s decision to side with Ellis meant that there would be equitable relief in a lower court even if the Ninth Circuit waited to adopt a version of the proposed rule until another case where the State opposes the prisoner’s habeas petition.[102]

Judge Owens, the other member of the Strickland camp who spoke at oral argument, articulated what would become the court’s solution to Ellis. He asked what would happen if the en banc panel simply acknowledged that the State was no longer advocating for a denial of habeas and remanded to the district court with instructions to grant habeas.[103]

2. The Cronic Camp

The judges in the Cronic camp echoed Judge Owens’s procedural mechanism for resolving Ellis even as they indicated that they believed a per se rule could be warranted—and indeed, eventually joined a concurrence saying as much. Judge Hurwitz explained that the district judge would look at the State’s position that Ellis should be let free because the court should presume prejudice and say “I’m not going to have to do anything ‘cause the [S]tate’s position is that we presume prejudice.”[104] Judge Nguyen anticipated that “[t]he district court would say [to the State], [‘]file the stipulation in a proposed order[’]”[105] i.e. that the State would agree to grant habeas without another trial. Judge Murguia inquired as to the mechanics of settling the case and how the court would proceed if it found that habeas was warranted. Chief Judge Thomas noted that overruling Mayfield and remanding is the “narrowest way we could approach reversal,”[106] but also suggested that he agreed with Nguyen’s underlying equitable principle granting habeas because “once there is either a conflict or you have a racial prejudice that’s severe, we can’t defer really to [the lawyer’s strategic] judgments because they are tinged.”[107]

Some might argue that the Ellis court “ducked” the hard problems in the case by simultaneously granting habeas but limiting the precedential impact of its decision to avoid the policy issues with “cabining” the per se rule articulated by Judges Smith and Hurwitz, discussed above. Rather than “ducking” the thorny doctrinal issues in Ellis, however, I believe the court’s decision to render an equitable result through procedural means underscores the ways in which doctrine constrained the court’s decision-making, even while the equities at stake compelled the court to find a solution that would grant Ellis habeas relief.

In Part III, I develop a theory of judicial decision-making that explains how and why courts retreat to the procedural core of their discretionary authority when deciding hard cases. I critique existing theories of judicial decision-making that either deny that such retreats occur at all or declare that these retreats occur in doctrinal environments without any normative constraints by the common law. Ellis suggests that when doctrinal and equitable factors counsel opposite results, courts use procedure to reach an equitable outcome without enacting precedent that the court sees as problematic.

III. Reconciling Legalism and Attitudinalism: Balancing Equity and the Law

This Part is an attempt to reconcile the dueling academic theories of legalism and attitudinalism that traditionally provide the framework for analyzing judicial decisions. Put simply, legalists believe that judges feel compelled to follow existing legal rules—whether they are derived from the constitution or precedent—rather than their own ideologies or policy preferences.[108] On the other hand, attitudinalists believe that judges base their legal decisions on policy preferences, with law serving only to rationalize what are essentially policy judgments.[109] Contrary to views of judicial decision-making that compel a choice between these models, I argue that legalism and attitudinalism explain the two essential modes of judicial decision-making—law and equity—and that judges decide cases through a blend of these two models.[110] Specifically, the interaction between legalistic and attitudinal factors, or law and equity, forms the discretionary space in which a judge can choose between valid case dispositions. Following Judge Posner, I refer to this area of judicial discretion as the “zone of reasonableness.”[111]

Further, I argue that while judges balance the vectors of law and equity differently based on their political ideology (and therefore perceive smaller or larger zones of reasonableness in a given case), legalistic concerns ultimately predominate by limiting the judicial avenues for expressing equitable concerns even where there is no threat of reversal by a higher court. In cases like Ellis, however, where equitable and doctrinal factors pull sharply in opposite directions, courts use procedural disposition—the core of their discretionary power—to arrive at equitable outcomes.

A. The Existing Debate: Legalism vs. Attitudinalism

In this Section, I give an overview of the conflict between legalism and attitudinalism and argue that each properly explains only a portion of judicial decision-making.

Legalists view judges as willing adherents to existing legal rules, regardless of the judges’ own ideologies and policy preferences. [112] As Judge Posner somewhat sarcastically puts it, “[t]he ideal legalist decision is the product of a syllogism in which a rule of law supplies the major premise, the facts of the case supply the minor one, and the decision is the conclusion.”[113] Under a legalistic conception of judicial decision-making, doctrine exerts a strong “gravitational force” on the judge’s decision-making, and there is little room for a judge to apply her own individual judgment regarding fairness, justice, and equity.[114] Legalists differ on how strong this gravitational force is,[115] and therefore on how much room should be left over for judges’ ideology to creep into their decision-making.[116] Legalism takes “‘the rule of law’ . . . not men” seriously, and remains the “judiciary’s ‘official’ theory of judicial behavior.”[117]

Legalism is not just about following precedent in the sense of “this case before us is very similar to precedential case X, so the result should be Y,” but also about maintaining the coherence of the entire legal regime.[118] For example, the discussion of a constitutional “case and controversy” requirement in Ellis is an example of a structural legal constraint. The nature of the common law itself means that these conditions are historically contingent—their boundaries are defined by the outcomes of canonical cases. But the judicial determination of whether the conditions are met is hardly mechanical. A full conception of legalism goes beyond adherence to precedential text and presupposes that the internalization of judicial norms creates within the judge a normative conception of their role that limits their exercise of judicial power.[119]

Empirical scholarship by legalist scholars who believe that “the law,” doctrinally defined, heavily influences judicial outcomes has shown that non-legal variables (e.g. ideology) explain far less variance in judicial decision-making than legal variables.[120] Specifically, in Decision Making in the U.S. Courts of Appeals, Frank Cross found “just one legal standard, affirmance deference to the lower court decision, is consistently significant statistically and by far the most important single variable substantively in explaining circuit court outcomes.”[121] Cross’s observation of the influence of precedent over and above judges’ ideology bolsters the legalist argument, especially because many cases are settled in the face of clear law “on the books” or are dealt with in unpublished opinions not included in empirical datasets.[122]

In contrast, Judge Posner and William Landes used a similar dataset to apply an attitudinal “rational-choice (economic) approach to judicial behavior.”[123] They attempted to show that leisure, self-expression (of political preferences), and esteem drive judicial decision-making.[124] Posner, Landes, and Lee Epstein continued to flesh out a labor economics model of judicial decision-making in The Behavior of Federal Judges: A Theoretical and Empirical Study of Rational Choice. Behavior of Federal Judges viewed judges as employees whose policy preferences are more important to them than their income and who are focused on job satisfaction.[125] One vexing conclusion was that so-called “group effects” where “judge[s] of one ideological persuasion will tend to go along with other judges on the panel even when they are of a different persuasion” were deemed to “mainly reflect effort aversion.”[126] In essence, Posner, Landes, and Epstein argued that judges acquiesce to a less desirable legal outcome to avoid doing work.[127]

Posner and Landes’ approach is typical of the attitudinal school of judicial behavior. Attitudinalists believe that policy preferences motivate judges’ legal decision-making, and law merely rationalizes such policy judgments.[128] Judges’ decisions, therefore, “directly correspond to their personal policy preferences.”[129] Under attitudinalism, judges are only prevented from implementing their policy preferences by strategic concerns (e.g., fear of reversal by a higher court that could lead to a legal outcome further from their policy preference).[130] For example, had there been a credible threat of appeal in Ellis, judges might have been reluctant to endorse a per se rule for fear that a more conservative Supreme Court might rule sharply in the other direction and declare that Cronic did not accommodate conflicts based on a lawyer’s racial animus.

There are two main threads of critique against attitudinalism. First, some modern attitudinal theories can also be explained by legalism. As one critic explained, the findings reported in Behavior of Federal Judges are “remarkably favorable to sophisticated versions of legalism.”[131] This is in part because the “judicial utility function” employed by Posner, Landes, and Epstein is “devoid of content that would enable the derivation of predictions about judicial decision-making.”[132] “The utility function and the labor economics model are consistent with any conceivable behavior by federal judges” because the preference structures of different judges (e.g., how much they value leisure) are unobservable.[133] Judges “could have identical ideology scores, but produce wildly different judicial decisions—because these differences can be explained by the judicial utility function on a post hoc basis.”[134] In other words, critics argue that attitudinal models are not falsifiable because the judicial preferences they are based on are revealed—more truthfully, constructed—by the decisions themselves.

A second common critique of attitudinalism is that it is “corrosive” to the judicial institution because it conceptualizes judges as politicians unmoored from the rule of law. This critique sees law under attitudinalism as “merely a cipher—a stand-in for the real motivations behind a [judicial] decision.”[135] Implicit in the attitudinal model is that judges are always seeking to expand their own power and to exert their policy preferences ever more effectively.[136] This critique, that judges are power-hungry and need to be constrained by “the rule of law” to prevent injustice via unbounded discretion, goes back at least as far as Blackstone’s critique of judges sitting in equity as “a loose cannon on the ship of state.”[137]

In modern parlance, equity sounds much like the “judicial activism” at the center of legalistic critiques of attitudinalism, as equity is defined as “the power to meet the moral standards of justice in a particular case” through mitigating “the rigidity of the application of strict rules of law.”[138] Modern defenders of equity note that “[c]ritics of equity have faulted its potential for variability . . . . Uncontrolled individualism on the bench, extremes of flexibility, might lead to judicial tyranny or the abnegation of all standards.”[139] English chancellors (equitable judges) in the 1600s did not publish official reports of their decrees or formally justify their decisions. They “viewed each case anew, unhindered by precedent when [they] wished.”[140]

Even today, equitable resolution of cases—even through procedural mechanisms—is less likely to create precedent. Rather, equitable decision-making creates bounded exceptions to existing legal regimes. If equity, boiled down, is merely the imposition of a judge’s judgment as to the “right” way to decide a case, then attitudinalism is merely a theoretical framework for the role of equity in modern courts. Conceptualizing attitudinalism as equity not only bridges a linguistic divide between political scientists and legal scholars, but also suggests an interplay between attitudinalism and law as a normative force that is at the bedrock of the common law tradition.

Viewed this way, attitudinalism and legalism describe two sides of the same coin. Attitudinalists recognize that jurisprudence is political. But in treating judges like senators, attitudinalists overlook the legal superstructure that restricts acceptable rationales for justifying a decision. Legalists recognize that doctrine constrains judicial decision-making but fail to account for the role intuitive notions of justice plays in shaping the outcome of a case and the justification for that outcome. In essence, “[l]aw matters, although how it matters to individual justices is highly variable.”[141] In the next two Sections, I first develop a model of judicial decision-making that views legalism and attitudinalism, or law and equity, as opposing vectors that define the acceptable zone of discretion in which a judge may decide a case and then apply that model to Ellis. This analysis reveals how judges turn to procedural mechanisms to allow for equitable resolution in difficult cases where doctrinal and equitable vectors pull sharply in opposite directions.

B. Developing a Model of Procedural Equity to Explain Cases Where Equitable and Doctrinal Factors are in Tension

In this Section, I sketch out a model of judicial decision-making that sees law and equity as two opposing forces. The result is a zone of acceptable judicial discretion where the judge can choose between valid case dispositions. Here, a disposition is not simply who wins, but also how the court justifies the case outcome. Recognizing that judicial behaviors, rather than a single behavior, govern judicial decision-making, I argue that legalism and attitudinalism (or law and equity) comprise the two principal modes of judicial decision-making where the zone of reasonableness is typically enlarged by equitable factors and constrained by legal ones. The balance of these forces determines the acceptable zone of reasonableness in a given case. This tradeoff between law and equity is influenced, but not determined, by the ideology of the judge. Different judges may put more or less weight on the equitable or legalistic factors in any given case, especially in cases where doctrinal and equitable considerations counsel different results. I call this method of adjudication “procedural equity.”

Accepting that attitudinalism describes equitable decision-making while also accepting that legalistic constraints have both normative and positive influence on judges’ decision-making in a given case, the next step is to theorize how these two modes of decision-making interact. To do this, we must consider the phenomenology of judging. That is, what it feels like for a judge to make a legal judgment. Following Wayne Martin, Posner argues that the act of judging makes one “conscious of both freedom and constraint.”[142] Freedom comes from the act of choosing between outcomes or, I argue, between ways of justifying an outcome where different justifications have different legal effects. Constraint arises because making a judgment is “a matter of deliberation . . . of weighing alternatives.”[143] Understanding legal judgment involves channeling a judge’s idea of how the case should come out to satisfy intuitive notions of justice (equitable or attitudinal decision-making) through the doctrinal framework presented to the judge by the facts of the case (the legalistic constraints) while acknowledging that judges “may have a variety of legal preferences” that inform their perception of legalistic constraints (e.g., how strong a hold does stare decisis have?), just as their equitable preferences inform the judge’s perception of the equities in the case.[144]

Legalistic forces are inherently constricting whether they are statutory, precedential, or based on a normative conception of the judge’s role (especially vis-à-vis politicians). These different types of legalistic forces may be restrictive in different ways, but they all channel a judge’s discretionary authority. Strickland, for example, functions as a doctrinal legalistic constraint. It requires judges, even those who strongly believe that the Strickland test is morally or equitably wrong, to follow the two-step analysis to accord with judicial norms and escape criticism from other judges who have internalized those norms.

Equitable or attitudinal forces, in contrast, may constrict or enlarge what a judge considers their “zone of reasonableness,” where they feel that they “ha[ve] discretion to decide a case either way without disgracing [them]self.”[145] For example, a judge like Judge Callahan—laser-focused on the question of factual innocence in Ellis—might perceive an already small discretionary space shrunken further by equities that, for her, tilt away from granting Ellis habeas relief. A judge like Judge Nguyen, on the other hand, who believes that racism’s pernicious effects are not likely to be visible on the cold trial record and that a per se rule of some kind is likely warranted, would find her discretionary space expanded by the equities in play.

Next, I apply this model to Ellis, where the Ninth Circuit chose an outcome that both parties already agreed on but did so in a way that was unlikely to produce a doctrinal precedential effect. Taking my analysis a step further, I argue that this move from doctrinal to procedural justifications for a decision—what I term “procedural equity”—is a generalized judicial response when precedent sharply constrains the zone of reasonableness, counseling a decision in favor of party A while equitable factors strongly counsel the opposite result in favor of party B. These types of cases have been characterized as “hard cases,” as in “hard cases make bad law,” because they present a choice between an unappealing—or doctrinally impermissible—precedent and a deeply unjust outcome for the litigants.

Ellis poses a difficult question for legal scholars who insist that courts always try to expand their own power and are only limited by external constraints. Contrary to the implicit assumption of the attitudinal model that judges are always seeking to exert their policy preferences,[146] here, nearly every member of the en banc panel balked at implementing the proposed rule.[147] This choice is made all the more curious by the absence of a threat of reversal in this case. If the court enacted the rule proposed by Ellis and the State, neither party would have been likely to appeal the decision as both parties would have been happy with the result.[148]

In Ellis, the en banc panel confronted a case where ten of the eleven judges believed that equity pointed towards habeas relief, but legalistic concerns about both the procedural posture of the litigants and the system health of the courts nonetheless predominated. Counter to attitudinal theories, liberal judges like Judge Hurwitz who appear to believe that, from a policy perspective, a per se rule is warranted nonetheless expressed legalistic reservations about justiciability. At the same time, conservative judges like Judges Smith and Bybee voted to grant habeas relief even as they refused to join either Nguyen’s concurrence providing legal justification for habeas relief or Watford’s concurrence arguing that the court could review Ellis’s case despite AEDPA. Counter to legalistic theories, the court searched for an equitable resolution to the case that granted relief while limiting Ellis’s precedential value. In effect, the court marshaled its procedural discretion to create an equitable result in the face of doctrinal constriction—in other words, “procedural equity.”

If the initial decision in Ellis denying habeas relief illustrated the legalistic constraints imposed on the court’s effort to grant relief for clients of racist lawyers, the concurrence and the subsequent en banc vote to rehear Ellis’s case represented the opening of an area of judicial discretion through procedural means in response to clear equitable concerns: how can a lawyer as racist as Ames effectively represent a member of the race he abhors? The en banc vote signifies that equitable concerns overwhelmed a result that was precedentially compelled by Mayfield and widened the zone of reasonableness to grant Ellis relief by other means. This opening of an area for discretion does not represent a shirking of judicial duty, as some have suggested,[149] but rather a pushback against legalistic constraints that compelled an outcome at odds with the judges’ equitable concerns.

While the State’s waiver of the Teague bar was essential for the proposal of the per se rule,[150] the State’s decision to support habeas relief increased the constricting legalistic pressure on the en banc court’s zone of reasonableness by calling into question the presence of a case or controversy. Simultaneously, however, the State’s support of Ellis’s petition meant that the court could pursue avenues other than adopting the per se rule to resolve the case while still ensuring Ellis obtained equitable relief. Even if the State refused to grant habeas on remand to the trial court, the Ninth Circuit would get another more justiciable bite at the apple. Put another way, the State’s decision to support habeas relief strengthened legalistic concerns by calling into question the presence of a case or controversy and weakened equitable ones by ensuring that the State would either grant habeas in the District Court or the Ninth Circuit would have another chance to hear the case. The result was a shrinking of the acceptable zone of reasonableness to its procedural core, absence of a new precedent, and equitable relief for Ellis.

C. Attorneys General in the Ninth Circuit Should Affirmatively Support Habeas Relief for Ames’s Clients and Clients of Other Racist Lawyers

I next explain why Ellis will likely have little precedential impact for criminal defendants represented by racist lawyers before noting how Ellis’s procedural precedent effectively delegates responsibility for habeas relief for other clients of racist lawyers to attorneys general.

First, under AEDPA, federal courts may only grant relief if the state court below either (1) makes a decision contrary to, or involving an unreasonable application of, clearly established federal law as established by the Supreme Court or (2) bases its decision on an unreasonable determination of the facts in light of the evidence presented in the state proceeding.[151] Because the Supreme Court “has not yet articulated the extent to which prejudice can be presumed from counsel’s extreme racism,”[152] “federal courts must defer to—even if they disagree with—a state court’s requirement that a habeas petitioner show a reasonable likelihood that an unbiased counsel would have produced a more favorable outcome.”[153] In Ellis, the Ninth Circuit was only able to review Ellis’s claim because the state court had applied a preponderance of the evidence standard, rather than a reasonable likelihood standard when evaluating Strickland’s second prong. Absent this mistake, the court “could not grant habeas relief because Ellis concedes that he cannot show prejudice under the general Strickland analysis.”[154]

Second, the Ellis court did not overrule Mayfield. Judge Nguyen’s concurrence explicitly noted as much, explaining that the three-judge panel previously concluded Mayfield overruled Frazer “insofar as Frazer’s analysis rested on Sullivan’s conflict analysis.”[155] But, Nguyen went on to explain, the court could rely on Frazer in the en banc concurrence because Frazer relied in some part on Cronic, which was not discussed in Mayfield.[156] So, although Nguyen’s Ellis concurrence rehabilitated and extended Frazer somewhat by applying the duty of loyalty concept to a case where Ames’s racism was not obvious on the trial record, Mayfield remains good law, and the Sullivan framework is still of no use for clients of racist lawyers.

Third, for a future decision to make doctrinal precedent, the court would require two conflicting conditions. First, the State must waive the Teague bar that ordinarily prevents courts from retroactively using new rules of criminal procedure as the basis for federal habeas relief.[157] And second, the State must oppose habeas to ensure that the case is justiciable to avoid concerns surrounding advisory opinions discussed during the en banc oral argument.[158] Threading this needle is unlikely, if not impossible: Why would a state oppose habeas but grant Teague in an attempt to create a per se rule?

These limitations leave Ellis with a conflicted legacy. On the one hand, Ellis goes further than perhaps any case before in extending habeas relief to clients of racist lawyers where racial animus is not apparent on the cold trial record. On the other, a per se rule of the kind articulated by both Ellis and California is unlikely to materialize. In a sense, Ellis appears to be an outer-bound on judicial relief for clients of racist lawyers.

There is one important precedential effect of Ellis that goes beyond doctrine, however. In summarily granting habeas because Ellis and the State agreed that habeas should be granted in this instance, the Ellis court implicitly created a procedural rather than doctrinal precedent. This procedural precedent can be used by attorneys general in the Ninth Circuit to grant habeas to other clients of racist lawyers, including Ames’s other clients. By stipulating to habeas relief in a district court, these attorneys general could grant new trials for clients of racist lawyers without having to thread the needle of conflicting criteria. In a sense, then, Ellis represents a shifting of discretionary authority from the courts to the state executive in adjudicating habeas for clients of racist lawyers. This allows the per se rule articulated by both California and Ellis to be implemented, but only in instances where prosecutors decide that habeas relief is warranted.

Conclusion

Examining Ellis in context with the prior three-judge panel opinion illuminates the pitfalls of conventional theories of judicial behavior. Legalism by itself cannot explain Nguyen’s three-judge concurrence, the subsequent successful en banc call, or the court’s eventual decision to grant habeas. The three-judge panel concluded that precedent under Mayfield required denying Ellis’s habeas petition, but that the equities in the case were so strong that such a denial “threatened to make a mockery of Ellis’s 6th Amendment right.”[159] The en banc court granted habeas relief without overruling Mayfield, even while it went further than many other courts in granting habeas relief without evidence of racial animus on the cold trial record. Just as legalism cannot explain why the en banc court granted relief without overruling binding precedent, legalism also cannot explain why two judges signed only the two-paragraph majority opinion without joining Wardlaw’s concurrence justifying the court’s review of Ellis’s Cronic claim. While attitudinalism can explain the behavior of some judges during the oral argument and the legal arguments supporting policy concerns, it cannot explain why judges whose policy preferences were in accordance with the rule balked during the oral argument at implementing it—citing legalistic concerns about justiciability. Nor can attitudinalism explain why judges who expressed serious policy concerns about the rule nonetheless voted to grant habeas. This interplay shows that law has internal normative forces and that legalistic concerns can limit judicial avenues for adjudicating disputes even when conventional constraints are absent.

Conceptualizing judicial decision-making as the output of two different judicial pathways—law and equity—helps us understand the court’s variable appetite for “occasional legislating.” In particular, this multidimensional model of judicial decision-making makes sense of the court’s reaction to the State’s decision to switch sides as enhancing the strength of legalistic concerns while opening equitable avenues for disposing of the case without creating a problematic precedent. While equity initially drove an expansion of the zone of reasonableness, the State’s decision to side with Ellis introduced new legalistic concerns which constrained the court’s options for deciding the case. Finally, in situations where equitable concerns drive an expansion of the zone of reasonableness but legalistic concerns constrain it, courts prefer procedural mechanisms for disposing of a case. It now falls on attorneys general within the Ninth Circuit to act on the procedural opening created by the Ninth Circuit and proactively support habeas for other clients of racist lawyers.

DOI: https://doi.org/10.15779/Z38VH5CJ86.

Copyright © 2021 Paul Messick, J.D., 2021, University of California, Berkeley, School of Law. I wish to thank Professors Rachel Stern, Rebecca Wexler, and Charles Weisselberg for their mentorship, feedback, and advice throughout the note writing process. Special thanks to Chelsea Hanlock, Joon Hwang, Debbie Oh, and Ayyan Zubair for their feedback, dialogue, and contributions. Thanks also to participants in Berkeley's Spring 2020 Student Note Workshop.

2. Ellis v. Harrison, 947 F.3d 555, 555–56 (9th Cir. 2020) (en banc). ↑

3. Richard A. Posner, How Judges Think 81 (2008). ↑

4. Unofficial Transcript of Oral Argument at 14, Ellis*, 947 F.3d 555 (No. 16-56188), 2018 WL 9522247 [hereinafter Transcript of Oral Argument]. ↑

5. Ellis, 947 F.3d at 555–56. ↑

6. Id. at 556 (“In light of the State’s concession that habeas relief is warranted, we summarily reverse the district court’s denial of Ellis’s petition.”). ↑

7. Id. at 574 (Callahan, J., dissenting). ↑

8. Id. at 557 (Nguyen, J., concurring) (“[T]he majority implicitly concludes that the state court’s opinion here ‘was contrary to, or involved an unreasonable application of, clearly established Federal law’ . . . .”). ↑

9. Because the State and Ellis agreed on the outcome of the case, neither party would appeal to the Supreme Court no matter the outcome in the Ninth Circuit. ↑

10. David L. Shapiro, Jurisdiction and Discretion, 60 N.Y.U. L. Rev. 543, 546 (1985). ↑

11. See Pauline T. Kim, Lower Court Discretion, 82 N.Y.U. L. Rev. 383, 388 (2007). ↑

12. Strickland v. Washington, 466 U.S. 668, 688 (1984). ↑

13. Id. at 687. ↑

14. United States v. Cronic, 466 U.S. 648, 659–660 (1984). ↑

15. Cuyler v. Sullivan, 446 U.S. 335, 348 (1980). ↑

16. Ellis v. Harrison, 947 F.3d 555, 559 (9th Cir. 2020) (en banc) (Nguyen, J., concurring) (quoting Strickland, 466 U.S. at 694). ↑

17. See Sheri Lynn Johnson, John H. Blume & Patrick M. Wilson, Racial Epithets in the Criminal Process, 2011 Mich. State L. Rev. 755, 768–72 (2011). ↑

18. State v. Davis, 872 So. 2d 250, 252 (Fla. 2004). ↑

19. Id. at 253. ↑

20. State v. Yarbrough, No. 17-2000-10, 2001 Ohio App. LEXIS 1930, at 15–16 (Ohio Ct. App. Apr. 30, 2001); see Johnson et al., supra* note 16, at 771–72. ↑

21. Yarbrough, 2001 Ohio App. LEXIS 1930, at *16. ↑

22. Jones v. Campbell, 436 F.3d 1285, 1304 (11th Cir. 2006). ↑

23. Johnson et al., supra note 16, at 770–71. ↑

24. Jones, 436 F.3d at 1304–05. ↑

25. United States v. Cronic, 466 U.S. 648, 658–59 (1984) (“There are, however, circumstances that are so likely to prejudice the accused that the cost of litigating their effect in a particular case is unjustified.”). ↑

26. Id. at 659. ↑

27. Frazer v. United States, 18 F.3d 778, 783, 785 (1994) (quoting Cronic, 466 U.S. at 660). ↑

28. Id. at 780. ↑

29. Id. at 783–84. ↑

30. Johnson et al., supra note 16, at 769; Cuyler v. Sullivan, 446 U.S. 335, 348 (1980). ↑

31. Mickens v. Taylor, 535 U.S. 162, 166, 171–72 (2002) (“[D]efects in assistance that have no probable effect upon the [case’s] outcome do not establish a constitutional violation.”). The conflict in Mickens, a capital case, was particularly severe as the counsel first represented the victim and then the alleged murder in unrelated, back-to-back cases. ↑

32. Mayfield v. Woodford, 270 F.3d 915 (9th Cir. 2001) (en banc). ↑

33. Id. at 924–25. ↑

34. Id. at 925. ↑

35. Id. at 940 (Graber, J., dissenting). ↑

36. Id. ↑

37. See Johnson et al., supra note 16, at 769; Mayfield, 270 F.3d at 925. ↑

38. See Johnson et al., supra note 16, at 769. ↑

39. Mayfield, 270 F.3d at 941 (Graber, J., dissenting). ↑

40. Id. ↑

41. Id. at 925 (majority opinion) (“None of the declarations [of racial epithets] alleged that Ames used racial epithets to describe Mayfield or that Ames’s alleged prejudice affected his representation of Mayfield.”). ↑

42. See, e.g., Transcript of Oral Argument, supra note 3, at 30–31; Mayfield*, 270 F.3d at 932–33. ↑

43. Ellis v. Harrison, 947 F.3d 555, 557–58 (9th Cir. 2020) (en banc) (Nguyen, J., concurring). ↑

44. See cases discussed supra Parts I.B. and I.C. ↑

45. See Ellis v. Harrison, 891 F.3d 1160, 1162 (9th Cir. 2018) (per curiam). ↑

46. See id. at 1166 (“We are therefore bound under Mayfield to reject [Ellis’s] claim.”). ↑

47. Ellis, 947 F.3d at 555–56 (“In light of the State’s concession that habeas relief is warranted, we summarily reverse the district court’s denial of Ellis’s petition.”). ↑

48. Id. at 556–64 (Nguyen, J., concurring). ↑

49. *Ellis v. Harrison, 891 F.3d 1160, 1162–63 (9th Cir. 2018) (per curiam), rev’d and remanded en banc*, 947 F3d. 555 (9th Cir. 2020). ↑

50. Id. at 1163. ↑

51. Id. ↑

52. Id. at 1163, 1166. ↑

53. Id. at 1163 n.1. ↑

54. Id. (noting the Ninth Circuit three-judge panel could not consider this updated declaration in deciding Ellis’s case “because the state courts had no opportunity to do so”). ↑

55. See Ellis, 974 F.3d at 556. ↑

56. Ellis, 891 F.3d at 1166 (“Ellis fails to identify any acts or omissions by Ames that ‘fell below an objective standard of reasonableness.’ We are therefore bound under Mayfield to reject his claim.” (citations omitted)). ↑

57. Id. at 1166–67 (Nguyen, J., concurring) (“Had we not been bound by Mayfield, I would have granted Ellis’s petition.”). ↑

58. Id. ↑

59. Id. at 1167. ↑

60. Id. ↑

61. U.S. Cts. for the 9th Cir., Ninth Circuit En Banc Procedure Summary (2017), http://cdn.ca9.uscourts.gov/datastore/general/2017/02/10/En_Banc_Summary.pdf [https://perma.cc/CLQ6-H24K] (“If a majority of the active, non-recused judges vote in favor of rehearing en banc, then the case is reheard by the en banc court.”). ↑

62. See id. ↑

63. See Ellis v. Harrison, 947 F.3d 555, 568 n.1 (9th Cir. 2020) (Callahan, J., dissenting) (noting that the State opposed habeas on various grounds, including an argument that Ellis could not prove that Ames’s racism negatively affected his representation). ↑

64. Transcript of Oral Argument, supra note 3, at *13. ↑

65. Some commentators have speculated that Kamala Harris’s record as attorney general hampered her presidential bid and influenced Xavier Becerra to change the State’s position. See Shaun Martin, Ellis v. Harrison (9th Cir. - Jan. 15, 2020), Cal. App. Rep. (Jan. 15, 2020), http://calapp.blogspot.com/2020/01/ellis-v-harrison-9th-cir-jan-15-2020.html [https://perma.cc/TV4C-5YAD]. ↑

66. *Transcript of Oral Argument, supra note 3, at 8. ↑

67. Id. at *10. ↑

68. Id. at 2, 8, 11, 14. ↑

69. Posner, supra note 2 at 78. ↑

70. California waived “the Teague bar exhaustion and any other procedural defenses with respect to the Sixth Amendment claim.” Transcript of Oral Argument, supra note 3, at 8. The waiver of the Teague bar was necessary for the court to retroactively apply any new rule to Ellis’s case because Teague bars the retroactive application of a newly announced rule of criminal procedure on collateral review. See *Teague v. Lane, 489 U.S. 288, 299–300 (1989). ↑

71. See U.S. Ct. of Appeals for the 9th Cir., 16-56188 Ezzard Ellis v. C. Harrison, YouTube (June 18, 2019), https://www.youtube.com/watch?v=Xobll1caqrA [https://perma.cc/UT72-DUDP] [hereinafter Video of Oral Argument]. ↑

72. See Transcript of Oral Argument, supra note 3, at 29; Video of Oral Argument, supra* note 70. ↑

73. See Transcript of Oral Argument, supra note 3, at 10–11; Video of Oral Argument, supra* note 70. ↑

74. Ellis v. Harrison, 947 F.3d 555, 556 (9th Cir. 2020). ↑

75. Id. at 564 (Watford, J., concurring). Outside of a limited number of exceptions, AEDPA forecloses federal habeas relief for any claim decided on the merits in state court. AEDPA imposes a highly deferential standard of review of state court decisions where habeas writs may only be issued if the state court’s denial is “contrary to, or involved an unreasonable application of, clearly established Federal law.” 28 U.S.C. § 2254(d)(1); see Ellis, 947 F.3d at 565 (Callahan, J., dissenting). In her dissent, Judge Callahan argues that all of Ellis’s claims have been decided on the merits in state court in accordance with clearly established federal law. Ellis, 947 F.3d at 565–67. Judge Watford’s concurrence argues that Ellis’s Cronic claim was never before the state court and AEDPA therefore does not bar review. Id. at 564 (Watford, J., concurring). ↑

76. Ellis, 947 F.3d at 556. (Nguyen, J., concurring). ↑

77. See Transcript of Oral Argument, supra note 3, at 23; Video of Oral Argument, supra* note 70 (“[T]he problem is, you know, under Strickland, for example, we defer to trial counsel’s strategic judgments. But once there is either a conflict or you have a racial prejudice that’s severe, we can’t defer really to those . . . judgments because they are tinged. [A]ll those thousands . . . of decisions that you make during a trial are now suspect because there’s no relationship, no loyalty to the client.”) (Thomas, J., speaking). ↑

78. See Transcript of Oral Argument, supra note 3, at 6; Video of Oral Argument, supra* note 70 (“I get your point, especially with respect to Mr. Ames. But you’re really talking about putting a bomb under the whole criminal justice system, to the degree you have a lawyer who has any alleged hint of racism.”) (Smith, J., speaking). ↑

79. See Transcript of Oral Argument, supra note 3, at 9; Video of Oral Argument, supra* note 70 (“[Y]ou’re asking us to do a case in which you have no disagreement about how it should turn out, to write what will be an advisory opinion for future cases because—and I think you’re correct on this, our prior case law is wrong. But that’s . . . a strange thing to be coming to a court to do, isn’t it?”) (Hurwitz, J., speaking). ↑

80. Transcript of Oral Argument, supra note 3, at 7; Video of Oral Argument, supra* note 70. ↑

81. *Transcript of Oral Argument, supra note 3, at 2; Video of Oral Argument, supra note 70. ↑

82. *Transcript of Oral Argument, supra note 3, at 17; Video of Oral Argument, supra note 70. ↑

83. *Transcript of Oral Argument, supra note 3, at 4; Video of Oral Argument, supra note 70. ↑

84. *Transcript of Oral Argument, supra note 3, at 6; Video of Oral Argument, supra note 70. ↑

85. See Transcript of Oral Argument, supra note 3, at 12; Video of Oral Argument, supra* note 70 (“Do you not have any relationship with the courts of California that you cannot get this—that you cannot get this rule adopted in California courts, without having to come to the federal courts and impose it on California?”). ↑

86. *Transcript of Oral Argument, supra note 3, at 8; Video of Oral Argument, supra note 70. ↑

87. *Transcript of Oral Argument, supra note 3, at 11; Video of Oral Argument, supra note 70. ↑

88. See Transcript of Oral Argument, supra note 3, at 6–7, 11; Video of Oral Argument, supra note 70. ↑

89. *Transcript of Oral Argument, supra note 3, at 10; Video of Oral Argument, supra note 70; see Wardlaw, Kim McLane, Fed. Jud. Ctr., https://www.fjc.gov/history/judges/wardlaw-kim-mclane [https://perma.cc/N4Y4-Y94F]. ↑

90. *Transcript of Oral Argument, supra note 3; Video of Oral Argument, supra* note 70. ↑

91. *Transcript of Oral Argument, supra note 3; Video of Oral Argument, supra* note 70. ↑

92. *Transcript of Oral Argument, supra note 3, at 24; Video of Oral Argument, supra note 70. ↑

93. *Transcript of Oral Argument, supra note 3, at 9; Video of Oral Argument, supra note 70. ↑

94. *Transcript of Oral Argument, supra note 3, at 3; Video of Oral Argument, supra note 70. ↑

95. *Transcript of Oral Argument, supra note 3; Video of Oral Argument, supra* note 70. ↑

96. See Transcript of Oral Argument, supra note 3 (“[W]hat if the rule is simply that when you . . . say racially derogatory things about your own client while you’re representing him . . . [t]hat’s enough to establish a conflict of interest. . . . [W]hy do we need a rule broader than that?”). ↑

97. Id. at 9; Video of Oral Argument, supra* note 70. ↑

98. *Transcript of Oral Argument, supra note 3; Video of Oral Argument, supra* note 70. ↑

99. See Kim, supra note 10, at 404 (“In the absence of [the threat of reversal], why would lower court judges—assumed to be motivated by their policy preferences—choose to follow legal authority rather than pursuing their own preferred outcomes?”). ↑

100. See Transcript of Oral Argument, supra note 3, at 10, 28–29. ↑

101. Id. at 9; Video of Oral Argument, supra* note 70. ↑

102. *Transcript of Oral Argument, supra note 3, at 1. ↑

103. Id. at 27; Video of Oral Argument, supra* note 70. ↑

104. *Transcript of Oral Argument, supra note 3, at 16; Video of Oral Argument, supra note 70. ↑

105. *Transcript of Oral Argument, supra note 3, at 16; Video of Oral Argument, supra note 70. ↑

106. *Transcript of Oral Argument, supra note 3, at 17; Video of Oral Argument, supra note 70. ↑

107. *Transcript of Oral Argument, supra note 3, at 23; Video of Oral Argument, supra note 70. ↑

108. See Barry Friedman & Andrew D. Martin, Looking for Law in All the Wrong Places: Some Suggestions for Modeling Legal Decision-Making, in What’s Law Got To Do With It?: What Judges Do, Why They Do It, and What’s at Stake 143, 153 (Charles Gardner Geyh ed., 2011) (noting that legalistic models of judicial behavior privilege are “an a priori commitment to a set of interpretive principles” rather than a commitment to certain policy outcomes). ↑

109. See Jeffrey A. Segal & Harold J. Spaeth, The Supreme Court and the Attitudinal Model Revisited 111 (2002) (arguing that the Supreme Court Justices may “freely implement their policy preferences”). ↑

110. In doing so, I expand on others’ observations that multiple judicial behaviors guide judicial decision-making, rather than a single behavior. See Stephen B. Burbank, On the Study of Judicial Behaviors: Of Law, Politics, Science, and Humility, in What’s Law Got To Do With It?, supra note 107, at 41, 47 (“[W]e should speak not of ‘judicial behavior’ but of ‘judicial behaviors.’”); Kim, supra note 10, at 385, 419–20. ↑

111. Posner defines the zone of reasonableness as “the area within which [the judge] has discretion to decide a case either way without disgracing himself.” See Posner, supra note 2 at 86. ↑

112. Friedman & Martin, supra note 107, at 152. ↑

113. See Posner, supra note 2, at 41. ↑

114. See Jeffrey A. Segal, What’s Law Got to Do with It: Thoughts from “The Realm of Political Science,” in What’s law Got To Do With It?, supra note 107, at 17, 18. ↑

115. Posner notes that legalism is usually conceptualized as what judges should be doing. See Posner, supra note 2, at 41–44. ↑

116. See Segal, supra note 113, at 18. ↑

117. See Posner, supra note 2, at 41. ↑

118. See Friedman & Martin, supra note 107, at 152–53 (discussing factors in the legal model that are focused on system coherence, such as plain meaning, precedent, and neutral rules). ↑

119. See Segal, supra note 113, at 18 (“[A] legal state of mind does not necessarily mean obedience to conspicuous rules; instead, it means a sense of obligation to make the best decision possible in light of one’s general training and sense of professional obligation.”). ↑

120. See Frank B. Cross, Decision Making in the U.S. Courts of Appeals 229 (2007) (“It is . . . noteworthy how very limited the explanatory power of the non-legal variables was . . . . After the law, the most consistently significant determinant of circuit court outcomes was ideology.”). ↑

121. Id. at 228. ↑

122. See id. at 48, 51, 207. ↑

123. See William M. Landes & Richard A. Posner, Rational Judicial Behavior: A Statistical Study 4 (U. Chi. L. & Econ., Olin Working Paper No. 404, 2008), https://papers.ssrn.com/abstract=1126403 [https://perma.cc/N4FY-GW43]. ↑

124. See id. at 5. ↑

125. See Lee Epstein, William M. Landes & Richard A. Posner, The Behavior of Federal Judges: A Theoretical and Empirical Study of Rational Choice 48 (2013). ↑

126. Id. at 192, 199. ↑

127. See id. at 199. ↑

128. For what is widely regarded as a modern touchtone of attitudinal theory, see Segal & Spaeth, supra note 108. ↑

129. Segal, supra note 113, at 26. ↑

130. See id. As we have seen, in Ellis some conventional constraints included in strategic models were not present, but judges still acted as though they were constrained. ↑

131. Lawrence B. Solum, The Positive Foundations of Formalism: False Necessity and American Legal Realism, 127 Harv. L. Rev. 2464, 2478 (2014) (reviewing Epstein et al., supra note 124). ↑

132. Id. at 2481. ↑

133. Id. ↑

134. Id. at 2482. ↑

135. Kim, supra note 10, at 385. ↑

136. Judge Posner has explained that this desire can also be ascribed to legalist judges who “will be inclined to declare a rule that will confer a broad precedential scope on the decision so that the next case, at least, can be decided deductively.” Posner, supra note 2, at 105. ↑

137. See Peter Charles Hoffer, The Law’s Conscience: Equitable Constitutionalism in America 11 (1990). ↑

138. Henry Lacey McClintock, Handbook of the Principles of Equity 1 (2d ed. 1948). ↑

139. Hoffer, supra note 136, at 17. ↑

140. Id. ↑

141. Michael A. Bailey & Forrest Maltzman, The Constrained Court: Law, Politics, and the Decisions Justices Make 16 (2011). ↑

142. Posner, supra note 2, at 85. ↑

143. Id. ↑

144. See Kim, supra note 10, at 404. ↑

145. Posner, supra note 2, at 86. ↑

146. See Kim, supra note 10, at 384. ↑

147. See supra Part II. ↑

148. See supra p. 112-13; n. 70-71. ↑

149. See, e.g., Donald R. Songer, Jeffrey A. Segal & Charles M. Cameron, The Hierarchy of Justice: Testing a Principal-Agent Model of Supreme Court–Circuit Court Interactions, 38 Am. J. Pol. Sci. 673, 693 (1994) (“[O]ur findings suggest that the appeals court judges were able to shirk, thereby partially advancing their own policy preferences, by interpretations of Supreme Court doctrine in ambiguous situations . . . .”). ↑

150. Ellis v. Harrison, 947 F.3d 555, 568 (9th Cir. 2020) (Callahan, J., dissenting) (“Acknowledging that its requested new rule would normally be barred on collateral review, the State expressly offered to waive the Teague bar . . . .”). ↑

151. 28 U.S.C. § 2254. ↑

152. Ellis, 947 F.3d at 560 (Nguyen, J., concurring). ↑

153. Id. at 560–61. ↑

154. Id. at 561. ↑

155. Id. at 561 n.8. ↑

156. Id. ↑

157. See id. at 560 n.6. ↑

158. *Transcript of Oral Argument, supra note 3, at 11–16; Video of Oral Argument, supra note 70. ↑

159. Ellis, 947 F.3d at 1166 (Nguyen, J., concurring). ↑