Mandala v. NTT Data, Inc., No. 19-2308 (2d Cir. 2021)Annotate this Case
This opinion or order relates to an opinion or order originally issued on September 21, 2020.
19-2308-cv Mandala v. NTT Data, Inc. UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT At a stated term of the United States Court of Appeals for the Second Circuit, held at the Thurgood Marshall United States Courthouse, 40 Foley Square, in the City of New York, on the 23rd day of February, two thousand twenty-one. Present: DEBRA ANN LIVINGSTON, Chief Judge, JOSÉ A. CABRANES, ROSEMARY S. POOLER, ROBERT A. KATZMANN, DENNY CHIN, RAYMOND J. LOHIER, JR., SUSAN L. CARNEY, RICHARD J. SULLIVAN, JOSEPH F. BIANCO, MICHAEL H. PARK, WILLIAM J. NARDINI, STEVEN J. MENASHI, Circuit Judges. _____________________________________ GEORGE MANDALA, CHARLES BARNETT, individually and on behalf of all others similarly situated, Plaintiffs-Appellants, v. 19-2308-cv 1 NTT DATA, INC., Defendant-Appellee. _____________________________________ For Plaintiffs-Appellants: Ossai Miazad, Lewis M. Steel, Christopher M. McNerney, Outten & Golden LLP, New York, NY; Rachel Bien, Outten & Golden LLP, Los Angeles, CA; Sherrilyn A. Ifill, Janai S. Nelson, Samuel Spital, Rachel M. Kleinman, NAACP Legal Defense & Educational Fund, Inc., New York, NY; Catherine Meza, NAACP Legal Defense & Educational Fund, Inc., Washington, DC. For Defendant-Appellee: Jacqueline Phipps Polito, Jessica F. Pizzutelli, Littler Mendelson P.C., New York, NY. Following disposition of this appeal on September 21, 2020, PlaintiffsAppellants filed a petition for rehearing en banc and an active judge of the Court requested a poll on whether to rehear the case en banc. A poll having been conducted and there being no majority favoring en banc review, the petition for rehearing en banc is hereby DENIED. Richard J. Sullivan and William J. Nardini, Circuit Judges, joined by Debra Ann Livingston, Chief Judge, and José A. Cabranes and Michael H. Park, Circuit Judges, concur by opinion in the denial of rehearing en banc. Rosemary S. Pooler, Circuit Judge, joined by Denny Chin, Raymond J. Lohier, Jr., and Susan L. Carney, Circuit Judges, dissents by opinion from the denial of rehearing en banc. Denny Chin, Circuit Judge, joined by Rosemary S. Pooler, Robert A. Katzmann, Raymond J. Lohier, Jr., and Susan L. Carney, Circuit Judges, dissents by opinion from the denial of rehearing en banc. 2 Raymond J. Lohier, Jr., Circuit Judge, joined by Rosemary S. Pooler, Robert A. Katzmann, Denny Chin, and Susan L. Carney, Circuit Judges, dissents by opinion from the denial of rehearing en banc. Peter W. Hall, Circuit Judge, took no part in the consideration or decision of the petition. FOR THE COURT: Catherine O’Hagan Wolfe, Clerk 3 RICHARD J. SULLIVAN and WILLIAM J. NARDINI, Circuit Judges, joined by DEBRA ANN LIVINGSTON, Chief Judge, and JOSÉ A. CABRANES and MICHAEL H. PARK, Circuit Judges, concurring in the order denying rehearing en banc: Unsurprisingly, we concur in the order denying rehearing en banc – we are, after all, the members of the majority that voted to affirm the district court’s dismissal of the complaint in this matter. We add this brief concurrence only to explain our belief that the dissents misapprehend the nature and consequences of the panel majority opinion, which reflects a heartland application of the plausibility pleading standard that has been the law of this Circuit for more than a decade. Put simply, we see no reason to fear that requiring Title VII plaintiffs to allege a plausible link between their chosen statistics and the qualified labor pool for the jobs in question will fundamentally alter the existing Title VII architecture. The thrust of the dissents’ argument is that statistics concerning the general population can be used to “nudge” a disparate impact claim “across the line from conceivable to plausible” at the pleading stage. Post, Chin, J., dissenting from denial of rehearing en banc, at 6 (internal quotation marks and brackets omitted); see also post, Pooler, J., dissenting from denial of rehearing en banc, at 3. As a general proposition, we agree. See Mandala v. NTT Data, Inc., 975 F.3d 202, 210–11 (2d Cir. 2020). But courts are not called on to announce general propositions; they are tasked with deciding particular cases based on specific pleadings. And the specific pleadings here do not plausibly allege that the general population is likely to be representative of the qualified labor pool for the jobs in question. In fact, the allegations in the complaint suggest that the general population is unlikely to be representative of the qualified labor pool. At the pleading stage, a Title VII disparate impact complaint must plausibly allege that (i) a specific employment practice or policy exists, (ii) a disparity exists, and (iii) there is a causal connection between the two. Id. at 207–09. While reference to statistics frequently satisfies this pleading burden, both caselaw and common sense make clear that not just any statistics will do. Id. at 209–11. After all, “statistics come in infinite variety and . . . their usefulness depends on all of the surrounding facts and circumstances.” Malave v. Potter, 320 F.3d 321, 327 (2d Cir. 2003) (quoting Watson v. Fort Worth Bank & Tr., 487 U.S. 977, 996 n.3 (1988)). In the disparate impact context, this means, among other things, that a plaintiff’s chosen statistics must focus on disparities between appropriate comparator groups – that is, the individuals holding the jobs at issue and “the qualified population in the relevant labor market.” Mandala, 975 F.3d at 210 (footnote omitted) (quoting Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 650 (1989), superseded by statute on other grounds, 42 U.S.C. § 2000e–2(k)). 2 Naturally, information about these particular groups may be difficult to obtain during discovery, let alone at the pleading stage. So we often allow plaintiffs to rely on surrogate statistics to prove disparities between comparator groups that they otherwise could not measure directly. In many cases, this includes statistics for the general population. Mandala, 975 F.3d at 210–11; see also Malave, 320 F.3d at 326. But not always. As the panel majority opinion concludes, general population statistics may be used only when there is reason to think that they will reflect the qualified labor pool for the positions in question. See Mandala, 975 F.3d at 211 (citing Dothard v. Rawlinson, 433 U.S. 321, 330 (1977); Malave, 320 F.3d at 326; and EEOC v. Freeman, 961 F. Supp. 2d 783, 798 (D. Md. 2013), aff’d, 778 F.3d 463 (4th Cir. 2015)); see also Hazelwood Sch. Dist. v. United States, 433 U.S. 299, 308 n.13 (1977) (explaining that “[w]hen special qualifications are required to fill particular jobs, comparisons to the general population . . . may have little probative value”). Not only is this rule consistent with precedent, it makes good sense. If there is no plausible link between the alleged disparate impact and the statistics that the plaintiff uses to plead his case, then those statistics are at most merely consistent with liability. And “[w]here a complaint pleads facts that are merely consistent with a 3 defendant’s liability, it stops short of the line between possibility and plausibility of entitlement to relief.” 1 Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009) (internal quotation marks omitted). Of course, as Judge Chin’s dissent indicates, and as the panel majority opinion acknowledges, “at the pleading stage, a plaintiff need not prove the accuracy of a statistical study’s findings or the rigor of its methodology; he need only generally allege the facts that, accepted as true, make his alleged injury plausible.” Post, Chin, J., dissenting from denial of rehearing en banc, at 13–14 (brackets omitted) (quoting John v. Whole Foods Mkt. Grp., Inc., 858 F.3d 732, 737 (2d Cir. 2017)); see also Mandala, 975 F.3d at 209–10 (citing John, 858 F.3d at 737). But one must be careful not to elide the distinction between testing the soundness of a statistical study, and determining whether that study and its statistics, if taken as true, make the plaintiff’s claim plausible. While John concerns the former, 858 F.3d The Supreme Court’s decision in Dothard v. Rawlinson does not counsel otherwise. To be sure, the plaintiffs in Dothard were permitted to rely on national height and weight data even though the defendants argued that the only relevant data was that of the Alabama-based pool of applicants for the corrections-officer position in question. 433 U.S. at 329–30; see also post, Chin, J., dissenting from denial of rehearing en banc, at 12. But that was only because “there was no reason to suppose that physical height and weight characteristics of Alabama men and women differ markedly from those of the national population.” Dothard, 433 U.S. at 330; see also Wards Cove Packing, 490 U.S. at 651–52 (rejecting the plaintiffs’ proffered statistics because there was no reason to believe that those statistics were representative of the qualified applicant pools for the positions in question). 1 4 at 737, the panel majority opinion turns on the latter. Put differently, John stands for the unremarkable proposition that, at the pleading stage, we accept as true the findings of statistical studies. But that does not mean that we must take as true every inference that a plaintiff asks us to draw from those findings no matter how attenuated. In affirming the dismissal of Plaintiffs’ complaint, the panel majority opinion reasoned that Plaintiffs had failed to provide any connective tissue between their proffered statistics and the qualified labor pool in question (indeed, just the opposite). The dissents assert that the panel majority opinion arrived at this conclusion by impermissibly drawing inferences against Plaintiffs. Not so. To start, Judge Chin’s dissent suggests that the panel majority opinion’s conclusion was premature because “the applicant pool . . . has not yet been defined,” post, Chin, J., dissenting from denial of rehearing en banc, at 23, and because we don’t yet know whether the “Salesforce developer” and “web developer” positions at issue here require specialized training or education not shared among the general population, id. at 24–25. 2 But Plaintiffs’ complaint says Judge Chin’s dissent also suggests that the panel majority opinion overlooks the fact that Plaintiffs’ claims apply to positions other than just Salesforce developer and web developer. Post, Chin, J., dissenting from denial of rehearing en banc, at 5, 24–25. But the complaint contains factual 2 5 otherwise. For one thing, “the positions’ titles alone . . . reflect that they require at least some educational or technical experience that is not shared by the general population.” Mandala, 975 F.3d at 211–12. For another, the complaint goes out of its way to highlight Plaintiffs’ educational and technical credentials – including Charles Barnett’s “Masters of Science in Computer Science Technology,” J. App’x at 14 – which indicates that those credentials are relevant to the jobs in question. Importantly, Plaintiffs never allege that those credentials are shared by the general population. Next, the dissents suggest that, even if the qualified labor pool in question is more educated than the general population, that does not prohibit Plaintiffs from pleading a plausible claim based only on general population statistics. Post, Chin, J., dissenting from denial of rehearing en banc, at 25–26; post, Pooler, J., dissenting from denial of rehearing en banc, at 3–7. Again, we disagree. Conviction rates and educational attainment are nearly certain to be inversely correlated on an absolute basis. See Mandala, 975 F.3d at 212 (reasoning that “it is not much of a stretch to imagine that arrest and conviction rates are negatively correlated with education (at least to some degree)”). This is not to allegations about only those two jobs; it does not identify any other positions that NTT offers. See Mandala, 975 F.3d at 212 n.6. 6 suggest any particular causal relationship between education and a decrease in conviction rates; there is simply a longstanding link between the two. In other words, the conviction rates for African Americans (and, for that matter, individuals of any race) will fall as we control for higher educational attainment. This conclusion is not, as Judge Pooler’s dissent suggests, “a negative inference [drawn] against the complaint that was not in the record.” Post, Pooler, J., dissenting from denial of rehearing en banc, at 4. “The determination of whether a complaint states a plausible claim for relief is ‘a context-specific task that requires the reviewing court to draw on its judicial experience and common sense.’” EEOC v. Port Auth. of N.Y. & N.J., 768 F.3d 247, 258 (2d Cir. 2014) (quoting Iqbal, 556 U.S. at 679). Here, common sense dictates that highly educated individuals can be expected to have lower conviction rates than the general population – indeed, Judge Pooler’s dissent acknowledges as much. See post, Pooler, J., dissenting from denial of rehearing en banc, at 5. The panel majority opinion simply applied that understanding in considering whether the allegations in the complaint stated a plausible claim for relief. In other words, the panel majority opinion simply concluded that Plaintiffs were asking the district court to draw inferences that, 7 based on common sense, were too attenuated from the supplied statistics to be plausible. Of course, as Judge Chin’s dissent points out, that absolute conviction rates will fall as we consider more highly educated segments of the population does not mean that the disparity between the conviction rates for African Americans and whites will necessarily disappear. Post, Chin, J., dissenting from denial of rehearing en banc, at 26. But what it does mean is that, based on the allegations in the complaint, we have no idea what the difference between African American and white conviction rates will be once we limit our focus to highly educated individuals. And that is precisely the point. All we know is that, for highly educated individuals of any race, their conviction rates are unlikely to look like the rates for the general population. Without more, then, the disparities observed in general population statistics are merely consistent with Plaintiffs’ disparate impact theory and do not make their claim plausible. 3 Judge Chin’s dissent asserts that the panel majority opinion simply invented this “confounding variable” rationale on its own. See post, Chin, J., dissenting from denial of rehearing en banc, at 4 n.1. Again, not so. That the general population statistics in question do not single-handedly render Plaintiffs’ claims plausible was addressed in the district court’s opinion, see Mandala v. NTT Data, Inc., No. 18-cv-6591 (CJS), 2019 WL 3237361, at *3 (W.D.N.Y. July 18, 2019) (reasoning that “[t]he statistics Plaintiffs cite in the complaint do not indicate whether the individuals in the general population cited shared qualifications that would make them viable candidates for either of the positions offered to Plaintiffs”), in NTT’s brief, see NTT Br. at 20–21, 21 n.10 (arguing that 3 8 As the panel majority opinion was careful to explain, this does not mean that national statistics can never be used in disparate impact cases involving skilled positions. Plaintiffs simply need to “provide additional allegations to explain why their chosen national statistics are in fact likely to be representative of [the] qualified applicant pool” in question. Mandala, 975 F.3d at 212. Here, that could have taken the form of additional national statistics indicating that, even as education levels increase, racial disparities between conviction rates remain. But Plaintiffs failed to provide such allegations. It is for that limited reason that the panel majority opinion affirmed the district court’s dismissal of Plaintiffs’ complaint for failure to state a claim. While that is sufficient to resolve the matter, it bears noting that new facts introduced by an amicus brief filed in support of rehearing the case confirm this conclusion and underscore the limited impact of the panel majority opinion. Specifically, a brief submitted by several criminology and sociology professors, Dkt. No. 146, identifies a study indicating that “Black men with some college education have imprisonment risks that are seven times greater than white men with some college education (4.9% for Black men compared to 0.7% for white “general population statistics have no application to Plaintiffs[’] claims” because “only qualified individuals” are subject to the at-issue policy), and at length during oral argument. 9 men),” post, Chin, J., dissenting from denial of rehearing en banc, at 9 (quoting Kurlychek Amicus Br. at 9); see also post, Pooler, J., dissenting from denial of rehearing en banc, at 6. In other words, the very figures that might have rendered Plaintiffs’ claims plausible not only exist but also are publicly available; Plaintiffs simply failed to include them in their pleadings. See Port Auth. of N.Y. & N.J., 768 F.3d at 258 (noting that “imprecise pleading is particularly inappropriate where the plaintiffs necessarily had access, without discovery, to specific information from which to fashion a suitable complaint” (internal quotation marks and alterations omitted)). As a result, we see no reason to believe that the panel majority opinion will become “a dangerous precedent” that shuts the courthouse door to disparate impact claims. Post, Chin, J., dissenting from denial of rehearing en banc, at 28. The fact that such publicly available statistics exist indicates that there is ample factual material on which future litigants may rely to plead plausible disparate impact claims. * * * To be clear, we do not take issue with the dissents’ descriptions of the significance of Title VII. See, e.g., post, Pooler, J., dissenting from denial of 10 rehearing en banc, at 7–8. Indeed, just the opposite. But even on matters of great importance, we are required to apply the pleading standards as set forth by the Supreme Court and this Court, and under those standards, Plaintiffs’ complaint falls short. For decades, Title VII claims – just like all other claims – were subject to a plaintiff-friendly notice pleading standard. See Swierkiewicz v. Sorema N.A., 534 U.S. 506, 512 (2002). That changed with the Supreme Court’s announcement of the plausibility pleading standard in Twombly and Iqbal. For better or for worse, that standard made it harder for all plaintiffs, not just Title VII plaintiffs, to state a claim for relief. Although one can surely debate the merits of this approach, neither the Supreme Court nor this Court has ever suggested that Title VII claims are somehow exempt from the plausibility standard. See Vega v. Hempstead Union Free Sch. Dist., 801 F.3d 72, 87 (2d Cir. 2015) (applying the plausibility standard to a Title VII disparate treatment claim); Littlejohn v. City of New York, 795 F.3d 297, 311 (2d Cir. 2015) (same). The panel majority opinion simply held Plaintiffs to that burden and agreed with the district court that they had failed to meet it in this particular case. Accordingly, we concur in the order denying rehearing en banc. 11 ROSEMARY S. POOLER, Circuit Judge, joined by DENNY CHIN, RAYMOND J. LOHIER, JR., and SUSAN L. CARNEY, Circuit Judges, dissenting from the order denying rehearing en banc: I join fully in Judge Chin’s thorough and compelling dissent from the order denying rehearing en banc. I write separately to emphasize two key issues. First, the panel opinion takes the wrong approach to Federal Rule of Civil Procedure 12(b)(6)’s plausibility standards by making inferences favoring the Defendants while declining to make obvious inferences for Plaintiffs that would rebut the central basis of the panel majority’s reasoning. Second, they take this flawed approach in the context of a Title VII lawsuit, undercutting one of the most important pieces of legislation in this country’s history. The principal flaw in the panel opinion is its inversion of the traditional standard applied to pleadings at the motion to dismiss stage. As our court has once before, this panel opinion has imposed a heightened standard for employment discrimination against the instructions of the Supreme Court. See Swierkiewicz v. Sorema N.A., 534 U.S. 506, 512 (2002) (“[I]mposing the Court of Appeals’ heightened pleading standard in employment discrimination cases conflicts with Federal Rule of Civil Procedure 8(a)(2).”). This case rests on a simple question—whether a blanket policy of excluding individuals with felony convictions from employment at NTT has a disparate impact on black applicants. To any person with a cursory understanding of America’s troubled racial history, the answer is clearly yes. Most Americans understand that the criminal justice system has quite clear racial biases that create disparate outcomes for black Americans. 1 These beliefs are accurate reflections of the state of criminal justice in this country. We should be clear. Black Americans are more likely to be arrested, convicted, and sentenced more harshly than are whites. In 2019, black Americans represented 13.4% of the U.S. population but 26.6% of all arrests by law enforcement, whereas whites represented about 76.3% of the population and 69.4% of arrests. 2 The panel opinion does not deny this but instead uses a statistical sleight of hand to hide the clear implications of NTT’s blanket policy. See, e.g., Drew Desilver, Michael Lipka & Dalia Fahmy, 10 Things We Know About Race and Policing in the U.S., Pew Rsch. Ctr. (Jun. 3, 2020), https://www.pewresearch.org/fact-tank/2020/06/03/10-things-we-know-aboutrace-and-policing-in-the-u-s/ (“Majorities of both black and white Americans say black people are treated less fairly than whites in dealing with the . . . criminal justice system as a whole. . . . 87% of blacks and 61% of whites said the U.S. criminal justice system treats black people less fairly.”) 2 See U.S. Census Bureau, QuickFacts: Population Estimates (July 1, 2019), https://www.census.gov/quickfacts/fact/table/US/PST045219; Fed. Bureau 1 2 Plaintiffs in this case offered general statistical studies showing the disparate conviction rates among the general population between black and white individuals. Joint App’x at 15. In Wards Cove Packing Co. v. Atonio, the Supreme Court recognized that “where figures for the general population might . . . accurately reflect the pool of qualified candidates, . . . we have . . . permitted plaintiffs to rest their prima facie cases on such statistics.” 490 U.S. 642, 651 n.6 (1989) (citations and internal quotation marks omitted), superseded by statute on other grounds, 42 U.S.C. § 2000e-2(k), as recognized in Texas Dep’t of Hous. & Cmty. Affs. v. Inclusive Cmtys. Project, Inc., 576 U.S. 519, 542 (2015). This is not a particularly difficult standard, and it seems clear that general population statistics “might” reflect the pool of qualified candidates here. At the motion to dismiss stage, it is certainly plausible that the general population statistics are reflective of the pool of applications to NTT. In response, the panel majority points to only one salient distinction between the general population and the applicant pool: education levels. See Mandala v. NTT Data, Inc., 975 F.3d 202, 211–12 (2d Cir. 2020). On this basis, the Investigation, Uniform Crime Reports, Arrests by Race and Ethnicity (2019), https://ucr.fbi.gov/crime-in-the-u.s/2019/crime-in-the-u.s.-2019/tables/table-43. 3 panel majority argues that the complaint failed to adequately allege that general population statistics are an appropriate proxy for the qualified applicant pool at issue here, reasoning that “it is not much of a stretch to imagine that arrest and conviction rates are negatively correlated with education (at least to some degree). So, while Plaintiffs’ statistics show that African Americans are on average more likely to have been convicted of a crime than whites, that does not, without more, make it plausible that an African American web developer with the educational and technical qualifications to work at NTT is more likely to have been convicted of a crime than his Caucasian counterpart.” Id at 212. This paragraph, critical to the panel majority’s reasoning, is where the majority takes a wrong turn. The panel majority explicitly acknowledges that it assumes what the data will show regarding education levels. In stating it is “not much of a stretch to imagine that arrest and conviction rates are negatively correlated with education (at least to some degree),” the majority draws a negative inference against the complaint that was not in the record. While statistics provided in amicus briefs reveal that the panel majority is likely correct that as education levels rise arrest and conviction rates fall, it is not clear why the panel majority considered it 4 appropriate to amend the record nostra sponte. Nor does the opinion concurring in the denial of rehearing en banc offer further insight as to where this determination may be found in the record. Instead, the opinion assures us that the court is “nearly certain” about this relationship, “there is simply a longstanding link between the two” factors, and “conviction rates for African Americans . . . will fall as we control for higher educational attainment.” Concurring Op. at 6. The concurrence concludes that this judicial determination was merely a matter of “common sense.” Id. at 7. I am quite willing to agree that education levels and conviction rates are likely negatively correlated. However, I am equally certain that racial disparities will continue to exist even after education levels are considered. It is “nearly certain” that this disparity will continue to exist, given the longstanding link between conviction rates and race across education levels, and conviction rates for black Americans remains higher than for whites as we control for higher educational attainment. While the concurrence argues that its assumptions regarding the relationship between education levels are a matter of “common sense,” that ability to use common sense and judicial experience to draw conclusions is absent once race enters the equation. In the next paragraph the 5 concurring judges profess ignorance of how race will interact with education, “[B]ased on the allegations in the complaint, we have no idea what the difference between African American and white conviction rates will be once we limit our focus to highly educated individuals. . . . All we know is that, for highly educated individuals of any race, . . . conviction rates are unlikely to look like the rates for the general population.” Concurring Op. at 8. Applying the same common sense that most Americans exercise in their views of the criminal justice system and our judicial experience overseeing that system, we know more than this. We know that racial disparities in conviction and arrest rates will persist across all education levels. Indeed, as set forth in another amicus brief, the data is quite clear on this point. See Brief for Megan C. Kurlychek et al. as Amici Curiae Supporting Appellants, Mandala v. NTT Data, Inc., 975 F.3d 202 (2d Cir. 2020) (No. 19-2308) at 8-9 (noting “[b]lack men with some college education have imprisonment risks that are seven (7) times greater than white men with some college education”). If the panel majority felt comfortable making its own assumptions regarding how education levels interacted with arrest and conviction rates, it is unclear why they did not feel comfortable making the equally obvious 6 assumption that racial gaps remain as education levels increase. I see no valid principle that permits the court to draw one inference but not the other, particularly at the motion to dismiss stage, all reasonable inferences must be drawn to favor plaintiffs. Our precedent clearly required the panel majority to either make both assumptions or neither, but the majority elected instead to employ its own standards to dismiss this case. The concurrence notes that the availability of statistics regarding the interaction between race, education, and conviction rates offer sufficient facts for future litigants to successfully plead plausible disparate impact claims. I encourage both future litigants to bring such cases and the Plaintiffs here to move under Rule 60 for relief from the district court’s judgment in order to file an amended complaint that includes statistics incorporating the continued racial gaps in conviction rates as education levels rise. While I hope the district court will allow such amendments and other courts will hear similar cases incorporating these statistics, it should not fall to litigants to correct for the panel majority’s failure to apply the proper standards. The flaws in the panel opinion are particularly important for our Court to remedy because they undercut Title VII. Title VII may be this century’s most 7 important piece of remedial legislation. Title VII struck a body blow to the racebased caste system that defined this country for centuries, and its promise of fair treatment has now thankfully been extended to the LGBT community. See Bostock v. Clayton County, --- U.S. ---, 140 S. Ct. 1731 (2020). By employing an eccentric and heightened pleading standard in this case, and importing facts from outside the record, the panel majority and concurrence suggest our Court will find ways to shut the door on litigants seeking to vindicate their civil rights. That is not the message we should send to litigants, especially in these troubled times. After months of protests, violence, and threats to the Nation’s most storied institutions and principles, more citizens than ever have questioned how different standards of treatment under law for black and white Americans have persisted from our founding to today. Instead of following our precedents and allowing this case to proceed for an examination of the consequences of these inequities, our Court has implemented a novel approach to shut the courthouse doors on plaintiffs. I respectfully dissent. 8 DENNY CHIN, Circuit Judge, joined by ROSEMARY S. POOLER, ROBERT A. KATZMANN, RAYMOND J. LOHIER, JR., and SUSAN L. CARNEY, Circuit Judges, dissenting from the order denying rehearing en banc: By denying the petition for rehearing en banc, the Court ignores a question of exceptional importance: the adverse impact of an absolute convictions bar on individuals seeking employment -- an impact disproportionately borne by African Americans. The heightened pleading standard created by the panel majority for disparate impact cases brought pursuant to Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq. ("Title VII"), presents a risk that many meritorious civil rights cases will not be reached on the merits. This is particularly troubling now in light of the implications for the struggle for racial equality that Title VII reflects, as the nation continues to address the issue of systemic racism. As the panel majority observes, "[f]acts are stubborn things," see Mandala v. NTT Data, Inc., 975 F.3d 202, 205 (2d Cir. 2020), and the statistics cited by plaintiffs in this case highlight an indeed "stubborn" -- and sobering -- fact: significant racial disparities in arrest, conviction, and incarceration rates persist in this country, disparities that have resulted in unfair and unnecessary barriers to employment. By acquiescing to the panel majority's use of a heightened pleading standard, the Court weakens Title VII, and threatens to "shut the courthouse door" to those who may have valid disparate impact claims. Petition for Rehearing at 16. As alleged in their complaint, plaintiffs-appellants George Mandala and Charles Barnett were offered jobs by defendant-appellee NTT Data, Inc. ("NTT") -- Mandala a position as a software consultant in New York and Massachusetts, and Barnett a position as a web developer in Kentucky. NTT withdrew the offers, however, after learning that Mandala and Barnett had previously been convicted of crimes. NTT did so pursuant to its blanket policy of denying employment to job applicants based solely on the fact of a prior conviction, without considering individual factors such as the nature and circumstances of the crime, its age, its bearing (if any) on the applicant's ability to perform the job sought, and evidence of rehabilitation and post-conviction good conduct. Mandala and Barnett, who are African American, brought this action, alleging that NTT's policy of rejecting applicants for employment because of prior convictions without individualized consideration has a disparate impact on African Americans seeking employment, in violation of Title VII. Plaintiffs 2 supported the allegations in their class action complaint with statistics showing that, on a national basis, "African Americans are arrested and incarcerated for crimes at higher rates than Whites, relative to their share of the national population." Joint App'x at 15. The district court granted NTT's motion to dismiss the complaint pursuant to Federal Rule of Civil Procedure 12(b)(6). The district court rejected plaintiffs' reliance on national statistics, holding that the statistics "do not indicate whether the individuals in the general population cited shared qualifications that would make them viable candidates for either of the positions offered to Plaintiffs." Joint App'x at 70. The panel majority affirmed. It largely adopted the district court's reasoning, holding that plaintiffs "provide no allegations to demonstrate that national arrest or incarceration statistics are in any way representative of the pool of potential applicants qualified for a position at NTT." 975 F.3d at 211. The panel majority rejected plaintiffs' reliance on national statistics, speculating that education is a "confounding variable" and musing that it is "not much of a stretch to imagine that arrest and conviction rates are negatively correlated with education (at least to some degree)." Id. at 211-12. The panel majority thus 3 presumes that the applicant pool for NTT is more highly educated than the nationwide population as a whole, and it surmises that racial disparities in arrest and conviction rates dissipate with increased education. Id. 1 To support its reasoning, the panel majority observes that "[a] simple example of this pitfall would be to apply national height averages to certain subgroups of the population, say NBA players and horse-racing jockeys." Id. at 211. But this case does not involve unique subgroups such as NBA players and horse-racing jockeys, elite athletes who indisputably are at opposite extremes of the height scale. Rather, the case involves two everyday individuals who received job offers for the not uncommon positions of software consultant and web developer, in different parts of the country, from a company with some 18,000 employees in North America and over twenty offices in the United States. Notably, the concept of a "confounding variable" formed the core of the panel majority's original decision. See 975 F.3d at 211. Yet, this was the panel majority's own creation. Indeed, even though NTT had access to data about its own qualified applicant pool, it did not argue in its brief on appeal that its pool of qualified applicants was more educated on average than the national population. See Appellee's Brief at 10-11 (arguing only that "qualified individuals" were at issue, that is, those NTT circularly defined as individuals who "applied for, were qualified for, and received an offer of employment with NTT"). Moreover, confounding variables are, by definition, variable. By declining to rehear the panel decision, the Court allows to stand a ruling that gives district judges discretion, before any discovery has been conducted, to select confounding variables that can proscribe the use of otherwise relevant general population statistics, and that permits appellate judges to inject such variables on their own initiative into the analysis on appeal. 1 4 While NTT's applicant pool may be different in some respects from the nationwide general population, it is certainly plausible that the workforce at NTT is not substantially different from the general population, and that any differences that do exist are not so significant as to render the national statistics irrelevant. The panel majority concludes otherwise -- even though the issue arose at the pleading stage, before any discovery, when information about NTT's applicant pools was not available. Even assuming, as the panel majority suggests, that the relevant applicant pool should be limited solely to the web developer and software consultant positions (which is likely unwarranted given the policy's blanket nature and geographic reach), it is certainly plausible -- even probable, as explained below -- that the racial disparity in conviction rates does not dissipate with education. And even if education serves to narrow the racial gap in conviction rates, there is no basis to conclude that education eliminates that disparity. In hypothesizing that NTT's applicant pool is so different from the general population that national arrest and conviction statistics must be inapposite, the panel majority draws an inference against plaintiffs, disregarding 5 governing pleading standards, controlling caselaw, and the well-settled principle that "[o]n a motion to dismiss, the question is not whether a plaintiff is likely to prevail, but whether the well-pleaded factual allegations plausibly give rise to an inference of unlawful discrimination, i.e., whether plaintiffs allege enough to 'nudge[ ] their claims across the line from conceivable to plausible.'" Vega v. Hempstead Union Free Sch. Dist., 801 F.3d 72, 87 (2d Cir. 2015) (citations omitted); accord Littlejohn v. City of New York, 795 F.3d 297, 310 (2d Cir. 2015). And in doing so, the panel majority imposes a heightened pleading standard for Title VII disparate impact cases, requiring something distinctly more than mere plausibility. For these and the further reasons discussed below, I respectfully dissent. I. The National Statistics As plaintiffs plausibly allege and national statistics show, "African Americans are arrested and incarcerated for crimes at higher rates than Whites, relative to their share of the national population." Joint App'x at 15. The complaint cites statistics from the Federal Bureau of Investigation (the "FBI"), the 6 U.S. Department of Justice, the U.S. Census Bureau, the U.S. Equal Employment Opportunity Commission (the "EEOC"), and scholars, as follows: as of 2010, 40% of prisoners in the United States were African American, while African Americans represented only 13% of the overall U.S. population (Prison Policy Initiative study) 2; some 26.9% of arrests are of African Americans, double their percentage of the general population (FBI and Census statistics); projections based on recent trends in incarceration estimate that one out of every three Black males born today will go to prison, compared to just one out of every seventeen White males, see Marc Mauer, Addressing Racial Disparities in Incarceration, 91 Prison J. 87S, 88S (2011); audit studies conducted by researchers at Harvard and Princeton found that African Americans with criminal records were particularly disadvantaged in the job market compared to Whites with criminal records, Joint App'x at 15 (citing scholarly journals); and the Department of Justice found that Blacks are arrested and convicted at higher rates than Whites, leading the EEOC to conclude that "[n]ational data . . . supports a finding that criminal record exclusions have a disparate impact based on race and national origin." Joint App'x at 15. The Prison Policy Initiative study reports racial disparities in incarceration rates not just nationally but in every state. See Leah Sakala, Breaking Down Mass Incarceration in the 2010 Census: State-by-State Incarceration Rates by Race/Ethnicity, Prison Pol'y Initiative (May 28, 2014) (https://www.prisonpolicy.org/reports/rates.html). 2 7 Department of Justice statistics, noted in the EEOC Enforcement Guidance, show that, as of 2010, Black men were incarcerated at almost seven times the rate of White men: Black men were imprisoned at the rate of 3,074 per 100,000, while White men were imprisoned at the rate of 459 per 100,000. 3 And much of this disparity cannot be attributed to the conduct of the individuals subjected to incarceration. For example, studies show that White and Black Americans are equally likely to use drugs and that White Americans are more likely to deal them; yet, Black Americans are arrested for drug crimes at far higher rates. See Am. Civil Liberties Union, A Tale of Two Countries: Racially Targeted Arrests in the Era of Marijuana Reform, at 28-29 (2020) (https://www.aclu.org/report/tale-two-countries-racially-targeted-arrests-eramarijuana-reform); Christopher Ingraham, White People Are More Likely to Deal Drugs, But Black People Are More Likely to Get Arrested For It, Wash. Post (Sept. 30, 2014) (https://www.washingtonpost.com/news/wonk/wp/2014/09/30/whitepeople-are-more-likely-to-deal-drugs-but-black-people-are-more-likely-to-get- See EEOC Enforcement Guidance on Consideration of Arrest and Conviction Records in Employment Decisions Under Title VII of the Civil Rights Act of 1964, (April 25, 2012) (https://www.eeoc.gov/laws/guidance/enforcement-guidance-considerationarrest-and-conviction-records-employment-decisions) (citing U.S. Dep't of Just., Bureau of Just. Stat., Prisoners in 2010, at 27 tbl.14 (2011) (https://bjs.gov/content/pub/pdf/ p10.pdf)). 3 8 arrested-for-it/); see also Kurlychek Amicus Brief at 11 ("[W]hite youth are more likely than minority youth to use marijuana, [but] studies continually show that minority youth are more likely to be arrested for such crimes, particularly marijuana possession."). While the panel majority speculates that the applicant pools for NTT are more highly educated than the general population and that racial disparities in arrest and conviction rates will decrease with education (drawing the inferences against plaintiffs), see 975 F.3d at 212, as amici point out, "[t]here are no available statistics to show that racial disparities in criminal conviction rates decrease, much less disappear, with increased education." Kurlychek Amicus Brief at 8-9. To the contrary, statistics show otherwise. For example, one study shows that "Black men with some college education have imprisonment risks that are seven times greater than white men with some college education (4.9% for Black men compared to 0.7% for white men)." Id. at 9. And even assuming the disparities lessen with increased education, there is no reason to assume that 9 they disappear altogether. Indeed, it is far more reasonable, at the pleadings stage, to infer that they do not. 4 Statistics also show that a criminal record has a substantially larger impact on Black Americans than on White Americans. See Devah Pager, The Mark of a Criminal Record, 108 Am. J. Soc. 937, 959 (2003) ("The effect of a criminal record is . . . 40% larger for blacks than for whites."); Devah Pager, Bruce Western, & Naomi Sugie, Sequencing Disadvantage: Barriers to Employment Facing Young Black and White Men with Criminal Records, 623 Annals Am. Acad. Pol. & Soc. Sci. 195, 196 (2009) (finding "a significant negative effect of a criminal record on employment outcomes, and one that appears substantially larger for African Americans"). Some states, including New York, recognize the harm that blanket criminal history screens can cause, forbidding companies from denying As amici note, Black Americans have disproportionately more contact with police officers than White Americans in the context of stop-and-frisks and traffic stops -situations where the educational level of the person being stopped or pulled over is not apparent. See Kurlychek Amicus Brief at 8-9. Amici also point to several examples of prominent, highly educated African Americans who have experienced disparate treatment based on race -- Professors Ronald S. Sullivan and Henry Louis Gates, Jr., of Harvard University and South Carolina Senator Tim Scott, who has reported that he has been pulled over by police seven times in one year. See id. at 8-10 & n.14 (citing Tanzina Vega, For Affluent Blacks, Wealth Doesn’t Stop Racial Profiling, CNN (July 14, 2016) (https://money.cnn.com/2016/07/14/news/economy/wealthy-blacksracialprofiling/index.html)). 4 10 employment solely because a job applicant has a criminal record, and instead requiring employers to engage in an individualized consideration. 5 As the Eighth Circuit recognized many years ago: We cannot conceive of any business necessity that would automatically place every individual convicted of any offense, except a minor traffic offense, in the permanent ranks of the unemployed. This is particularly true for blacks who have suffered and still suffer from the burdens of discrimination in our society. To deny job opportunities to these individuals because of some conduct which may be remote in time or does not significantly bear upon the particular job requirements is an unnecessarily harsh and unjust burden. Green v. Missouri Pac. R.R. Co., 523 F.2d 1290, 1298 (8th Cir. 1975). These observations still hold true today. The New York statute, for example, provides that an employer may not deny employment because of an individual's prior conviction unless (1) there is "a direct relationship" between the offense and the position sought or (2) the granting of employment would present an "unreasonable risk" to property, specific individuals or the general public. N.Y. Correct. Law § 752 (McKinney 2019). In making this determination, the employer "shall consider," inter alia, the "bearing, if any," the offense will have on the applicant's fitness or ability to perform his duties, the time elapsed since the conviction, the applicant's age at the time the offense was committed, and information concerning rehabilitation and good conduct. N.Y. Correct. Law § 753 (McKinney 2019); see also N.Y. Exec. Law § 290 et seq. (McKinney 2019); N.Y. Correct. Law § 750 et seq. (McKinney 2019). As a New York resident, Mandala asserted state law claims, but the district court did not reach them. 5 11 II. Plaintiffs Properly Relied on National Statistics As the majority acknowledges, national statistics can be probative of whether a challenged policy has a disparate impact. Mandala, 975 F.3d at 210. In Dothard v. Rawlinson, for example, the Supreme Court observed that "[t]here is no requirement . . . that a statistical showing of disproportionate impact must be based on analysis of the characteristics of actual applicants." 433 U.S. 321, 330 (1977). There, the plaintiffs were permitted to rely on national height and weight data for men and women, even though the defendant argued that the only relevant data was that of the applicant pool for corrections-officer positions in Alabama. Id. at 329-30. Indeed, the Supreme Court specifically rejected the defendant's argument that "a showing of disproportionate impact on women based on generalized national statistics should not suffice to establish a prima facie case," id. at 330, the very principle that forms the basis of the panel majority's holding here. Likewise, in Malave v. Potter, this Court rejected the district court's holding, on a summary judgment motion, that plaintiffs were required to provide statistical information as to "the applicant pool or the eligible labor pool." 320 F.3d 321, 325-26 (2d Cir. 2003). We rejected the district court's 12 "adoption of a rule that the lack of statistical information as to an applicant pool always renders it impossible to establish a prima facie disparate impact case." Id. at 327. We remanded for the district court to determine, inter alia, "the most appropriate labor pool," and reminded the district court of "the Supreme Court's teaching that 'statistics come in infinite variety and . . . their usefulness depends on all of the surrounding facts and circumstances.'" Id. (quoting Watson v. Fort Worth Bank & Tr., 487 U.S. 977, 995 n.3 (1988)); see also EEOC v. Joint Apprenticeship Comm. of Joint Indus. Bd. of Elec. Indus., 164 F.3d 89, 97 (2d Cir. 1998) (at summary judgment, plaintiff's prima facie case can include "studies based on general population data"). And in Wards Cove Packing Co. v. Atonio, the Supreme Court recognized that "where 'figures for the general population might . . . accurately reflect the pool of qualified candidates,' . . . we have even permitted plaintiffs to rest their prima facie cases on such statistics as well." 490 U.S. 642, 651 n.6 (1989) (citation omitted and emphasis added). In the context of a motion to dismiss, a plaintiff has even more latitude in relying on national statistics. We have held that "[a]t the pleading stage, [a plaintiff] need not prove the accuracy of [a statistical study's] findings or the rigor of its methodology; he need only generally allege the facts that, 13 accepted as true, make his alleged injury plausible." John v. Whole Foods Mkt. Grp., Inc., 858 F.3d 732, 737 (2d Cir. 2017); see also id. ("a facial attack on the pleadings is not the proper stage to determine whether [the Department of Consumer Affairs'] sampling methods justified its declaration of widespread overcharging"); Brown v. City of N.Y., No. 16 Civ. 1106, 2017 WL 1102677, at *6 (E.D.N.Y. Mar. 23, 2017) ("[S]tatistics that may ultimately prove insufficient can nevertheless support a plausible inference of disparate impact on a motion to dismiss."). Numerous courts have applied these principles to deny motions to dismiss disparate impact claims that rely on general population statistics to challenge such policies, concluding that plaintiffs plausibly stated a disparate impact claim under Title VII. 6 And the EEOC has specifically found that national See, e.g., Chaidez v. Ford Motor Co., 937 F.3d 998, 1007 (7th Cir. 2019) ("'Disparate-impact plaintiffs are permitted to rely on a variety of statistical methods and comparisons to support their claims,'" and "'at the pleading stage, some basic allegations of this sort will suffice.'" (citation and alterations omitted)); Smith v. Home Health Sols., Inc., No. 17 Civ. 30178, 2018 WL 5281743, at *3 (D. Mass. Oct. 24, 2018) (denying motion to dismiss disparate impact claim where plaintiffs alleged African Americans "are arrested in numbers disproportionate to their representation in the general population," citing "national data"); Williams v. Compassionate Care Hospice, No. 16 Civ. 2095, 2016 WL 4149987, at *4-5 (D.N.J. Aug. 3, 2016) (denying motion to dismiss disparate impact claim where plaintiffs alleged "blanket criminal background check" led to a "disproportionate impact on African Americans," relying on national statistics); 6 14 data "supports a finding that criminal record exclusions have a disparate impact based on race and national origin." EEOC Enforcement Guidance, supra. Here, plaintiffs did not rely solely on general population statistics nor did they rely on statistics in a conclusory or abstract fashion. Rather, they alleged specific facts and circumstances showing that NTT's blanket convictions bar had an adverse impact on them personally. Mandala applied for a salesforce developer position while residing in Rochester, New York, and he was offered a position as an application software development senior principal consultant in Wellesley, Massachusetts (to work remotely). Barnett applied for a web developer position while residing in Williams v. Wells Fargo Bank, N.A., No. 4-15-cv-38, 2015 WL 13753220, at *1 (S.D. Iowa Aug. 6, 2015) (denying motion to dismiss claim brought by seven African American plaintiffs alleging that bank's hiring and background check policy "disproportionately affected minorities"); McCain v. United States, No. 14 Civ. 92, 2015 WL 1221257, at *17 (D. Vt. Mar. 17, 2015) (denying motion to dismiss disparate impact claim challenging "blanket" background check policy, relying on national statistics); see also Lee v. Hertz Corp., 330 F.R.D. 557, 561 (N.D. Cal. 2019) ("Since Plaintiffs aver Latinos were arrested and convicted of crimes at more than double the rates of whites, it is plausible that Defendants' Screening Policy has a disparate impact on Latinos by tending to deprive them of employment opportunities because of their race or national origin."); Houser v. Pritzker, 28 F. Supp. 3d 222, 248-49 (S.D.N.Y. 2014) (certifying class of applicants denied employment because of criminal histories due to blanket criminal screen); Gregory v. Litton Sys., Inc., 316 F. Supp. 401, 403 (C.D. Cal. 1970) (criminal arrest screen "has the foreseeable effect of denying black applicants an equal opportunity for employment"), aff'd as modified, 472 F.2d 631 (9th Cir. 1972). 15 Frankfort, Kentucky, and was offered employment designing websites for the Kentucky Department of Education. Both received offers of employment from NTT, but had their offers withdrawn once NTT learned that they had prior convictions. NTT never asked Mandala and Barnett for information about their convictions, rehabilitation, or post-conviction conduct. Barnett, for example, had obtained two degrees, including a master's in computer science technology -after his conviction. He also worked for four years for the Commonwealth of Kentucky doing technology and administrative work -- after his convictions. NTT did not consider these post-conviction developments because of its blanket policy. Mandala and Barnett were clearly qualified for the NTT positions -they were offered employment -- but the policy barred them from employment without any individualized consideration of the circumstances of their convictions, the relationship between their criminal history and their ability to perform the jobs, or their efforts to rehabilitate and post-conviction conduct. Both were impacted by NTT's policy even though they lived in different parts of the country and applied for different jobs. After Barnett's offer was withdrawn, he "sought to apply for other positions with NTT," but NTT would not consider 16 him for any position because of its policy. Joint App'x at 14. Mandala and Barnett are striking examples of the adverse impact a blanket convictions bar can have on individuals and their families. In addition to these specific individual allegations, the complaint asserted broader allegations: NTT's policy "systematically eliminates qualified African American applicants based on their race, color or national origin" throughout the United States, Joint App'x at 8, and NTT is a "global" information technology services company with some 18,000 employees in North America and over twenty offices in the United States. In light of the breadth of NTT's blanket policy -- it applies to all jobs, at all levels, all across the country -- plaintiffs' reliance on national statistics at the pleadings stage was eminently reasonable, and the national statistics surely made plaintiffs' claims of disparate impact discrimination even more plausible. III. The Panel Majority Misapplied the Pleading Standards On a disparate impact claim, a plaintiff is required only to prove that "a facially neutral employment policy or practice has a significant disparate impact" on members of a protected group -- the plaintiff need not prove intent to discriminate. Brown v. Coach Stores, Inc., 163 F.3d 706, 712 (2d Cir. 1998); see Ricci 17 v. DeStefano, 557 U.S. 557, 577-78 (2009); Griggs v. Duke Power Co., 401 U.S. 424, 431 (1971) (Title VII prohibits "not only overt discrimination but also practices that are fair in form, but discriminatory in operation"); 42 U.S.C. § 2000e2(k)(1)(A)(i). If the plaintiff shows that a facially neutral employment policy has a disparate impact and thereby "establishes a prima facie violation," the employer may defend by demonstrating that the policy is job-related and consistent with business necessity. See Ricci, 557 U.S. at 578. If the employer does so, the plaintiff can still prevail by showing that the employer's legitimate interests can be served by an alternative that has less of a disparate impact. See id. At the pleadings stage, a plaintiff is required only to allege facts giving rise to a plausible inference of a disparate impact based on race -- a plausible inference that an employment practice "has the effect of denying members of a protected class equal access to employment opportunities." M.O.C.H.A. Soc'y, Inc. v. City of Buffalo, 689 F.3d 263, 273 (2d Cir. 2012). On a motion to dismiss for failure to state a claim, "the question is not whether a plaintiff is likely to prevail, but whether the well-pleaded factual allegations plausibly give rise to an inference of unlawful discrimination, i.e., whether 18 plaintiffs allege enough to 'nudge[ ] their claims across the line from conceivable to plausible.'" Vega v. Hempstead Union Free Sch. Dist., 801 F.3d 72, 87 (2d Cir. 2015) (citations omitted); see also Swierkiewicz v. Sorema N.A., 534 U.S. 506, 511 (2002) ("under a notice pleading system, it is not appropriate to require a plaintiff to plead facts establishing a prima facie case"). 7 Until now, we have consistently recognized that at the pleadings stage, without discovery, a proper record usually has not yet been developed. Accordingly, the plausibility standard "simply calls for enough fact to raise a reasonable expectation that discovery will Our cases have held that a plaintiff in an employment discrimination case has only a "'minimal burden' of alleging facts 'suggesting an inference of discriminatory motivation.'" Vega, 801 F.3d at 85 (quoting Littlejohn v. City of New York, 795 F.3d 297, 310 (2d Cir. 2015)). The panel majority contends that this "minimal burden" language applies only to disparate treatment cases and not to disparate impact cases. Mandala, 975 F.3d at 209 n.3. First, this novel legal view was first raised by the panel majority, as neither party argued that the pleading burden for Title VII cases depends on whether a claim alleges disparate impact or disparate treatment discrimination. See Appellants' Brief at 16-17 (arguing that "this Court has repeatedly rejected attempts to heighten the pleading standard in discrimination cases" (emphasis added)); Appellee's Brief at 13-14 (citing to two disparate treatment cases -- to argue that "[i]n the context of a Title VII claim," to survive a motion to dismiss, a plaintiff must plead facts that plausibly support a prima facie case (citations omitted)). Second, there is nothing in the language of Title VII to suggest that disparate treatment and disparate impact cases should be treated differently in the way the panel majority contemplates. Hence, I am not persuaded that there is any basis to distinguish in this respect between Title VII disparate treatment claims and Title VII disparate impact cases. See Petition for Rehearing at 8-9. While I do not believe that whether the "minimal burden" language applies to disparate impact cases makes a difference here, again, by declining to hear the case en banc the Court dodges an opportunity to resolve an important issue. 7 19 reveal evidence of illegal[ity]." Arista Records, LLC v. Doe 3, 604 F.3d 110, 120 (2d Cir. 2010) (quoting Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 570, 556 (2007)); accord Citizens United v. Schneiderman, 882 F.3d 374, 380 (2d Cir. 2018) (for plaintiffs to "'nudge[ ] their claims across the line from conceivable to plausible,' they must 'raise a reasonable expectation that discovery will reveal evidence' of the wrongdoing alleged, 'even if it strikes a savvy judge that actual proof of these facts is improbable'") (quoting Twombly, 550 U.S. at 570, 556)). 8 Here, the district court misapplied the standards applicable to Rule 12(b)(6) motions to dismiss. Although it acknowledged the "plausibility standard" in discussing the "general legal principles," Joint App'x at 67-68, in determining whether the complaint stated a plausible claim, the district court relied on a decision setting forth the standards for proving -- not pleading -- a There is, of course, a lack of precision as to what is "plausible" for purposes of Federal Rule of Procedure 12(b)(6), and different panels of this Court have interpreted the standard in different ways. By denying the petition for rehearing en banc, the Court also misses an opportunity to resolve whatever intra-circuit conflicts exist among these decisions regarding the standard. What is clear is that the plain meaning of "plausible" is, inter alia, appearing worthy of belief, or superficially fair or reasonable, or credible -not "probable" or likely to succeed. See, e.g., Plausible, Merriam-Webster Dictionary, https://www.merriam-webster.com/dictionary/plausible?src=search-dict-hed. The fact that several of my colleagues and I believe that plaintiffs have plausibly asserted a claim of disparate impact discrimination would suggest that, indeed, the claim of disparate impact discrimination is credible, at least superficially fair or reasonable, and crosses the line from "conceivable to plausible." 8 20 disparate impact case, Joint App'x at 69 (citing Robinson v. Metro-North Commuter R.R., 267 F.3d 147 (2d Cir. 2001)). The district court then rejected plaintiffs' reliance on national statistics, holding that plaintiffs' national statistics did not "indicate whether the individuals in the general population cited shared qualifications that would make them viable candidates for either of the positions offered to Plaintiffs." Joint App'x at 70. The district court held that "general statistics are inadequate," and that plaintiffs were required instead to provide statistics as to conviction rates with respect to "the pool of applicants who are Caucasian versus African Americans," Joint App'x at 70-71, the precise statistical evidence they would have to offer to prove their claim at trial. The district court reached these conclusions without identifying the relevant applicant pool, and before plaintiffs had an opportunity to take discovery. By affirming the district court's decision, the panel majority adopts the district court's deeply flawed reasoning that plaintiffs were required to provide, at the pleadings stage and before discovery, statistics as to conviction 21 rates for NTT's specific applicant pool. 9 In doing so, the panel majority makes numerous fundamental errors: First, it rejects plaintiffs' reliance on national statistics, when numerous courts -- including this Court and the Supreme Court -- have held that national statistics can support a disparate impact claim. At the pleadings stage in particular, plaintiffs are to be given some latitude in relying on statistics. See, e.g., John, 858 F.3d at 737. Second, the panel majority ignores the likelihood that, even if specific applicant pool statistics turn out to be more precise in the end, general population statistics would still be relevant and a logical starting point for the analysis. See Williams, 2016 WL 4149987, at *5 ("Although the Court agrees with Defendant that the [national] statistics are not tailored to the New Jersey counties See, e.g., 975 F.3d at 206 ("Notably, however, the complaint contains no allegations about racial disparities in NTT's existing workforce or the demographics of qualified applicants that NTT has rejected as a result of its hiring policy."), 211 ("Plaintiffs provide no allegations to demonstrate that national arrest or incarceration statistics are . . . representative of the pool of potential applicants qualified for a position at NTT."), 212 (Plaintiffs did not provide statistics "mak[ing] it plausible that an African-American web developer with the educational and technical qualifications to work at NTT is more likely to have been convicted of a crime than his Caucasian counterpart." (emphasis added)). 9 22 in which Defendant does business, the Court finds that they lend sufficient support to Plaintiff's allegations to survive the Motion to Dismiss."). Third, the majority holds that plaintiffs should have provided statistics pertaining to NTT's applicant pool -- "more granular data," Mandala, 975 F.3d at 212 -- when the applicant pool (assuming the relevant applicant pool is something other than the national workforce) has not yet been defined and plaintiffs have not yet had discovery. Even assuming that statistics as to a specific applicant pool, e.g., salesforce developers in upstate New York, software consultants in Massachusetts, web developers in Kentucky, or -- in the majority's words -- applicants "representative of the pool of potential applicants qualified for a position at NTT," id. at 211, turn out to be more accurate in the end, the absence of such statistics at the motion-to-dismiss stage should not be fatal to plaintiffs' claims, as the appropriate applicant pool cannot be defined until after discovery, when more details about NTT's job requirements and applicant pools would become available. Indeed, as the panel majority acknowledges, see id. at 212 (recognizing that plaintiffs are "undoubtedly working from an informational 23 disadvantage at this early point in the proceedings"), some of this information surely is only in NTT's possession and not publicly available. 10 Fourth, the panel majority fails to acknowledge the reasonable likelihood that the relevant applicant pool is the national workforce, given that plaintiffs sought employment with NTT in different parts of the country and for different positions, NTT is a global company, with some 18,000 employees and twenty offices in the United States, and its blanket policy applies to all jobs, at all levels, all across the country. The panel majority focuses on "salesforce developer" and "web developer" positions, jobs it speculates "require at least some educational or technical experience that is not shared by the general population." Id. at 211-12. It is not apparent, however, that "salesforce developer" and "web developer" positions, whatever they may be, require such specialized training or education as to make general population statistics Discovery would shed light on, for example, the following questions: what were the positions in question? what were the qualifications necessary for those positions? what was "the pool of qualified candidates?" what was the appropriate geographic area to be considered? was NTT considering applicants on a national basis because successful candidates would work remotely? what are the contours of NTT's criminal history screen? what were the demographics of the individuals in the pool of qualified candidates? what were the demographics of those who applied? were there individuals who did not apply because of the blanket policy? and what were their demographics? 10 24 inapplicable. Moreover, the panel majority's focus on the two positions ignores plaintiffs' allegations that after Barnett's offer was rescinded for the "web developer" position, he "sought to apply for other positions with NTT" but was told he would not be considered for "other positions because of his felony convictions." Joint App'x at 13-14. Plaintiffs' claims are not limited to salesforce developer and web developer positions, but the panel majority chooses simply to ignore this fact. Fifth, the panel majority speculates that NTT's applicant pool is more educated than the national population, when it is plausible, given NTT's size and geographic reach, that any differences in education levels are insignificant. Sixth, the panel majority assumes the existence of a "confounding variable" and "imagine[s]" that "arrest and conviction rates are negatively correlated with education (at least to some degree)," without acknowledging the likelihood that racial disparities will continue to exist to some degree within the qualified applicant pool, even as the level of education increases. 975 F.3d at 21112. While the panel majority's assumptions would mean that the qualified applicant pool would have lower conviction rates, those assumptions would not 25 necessarily eliminate, or even reduce, the disparities in conviction rates within the pool, that is, disparities based on race between similarly educated groups. Even assuming that NTT's applicant pool is more educated than the general population, it is unlikely that the racial disparities in arrest and incarceration rates that exist nationally and in every state somehow disappear at NTT. Seventh, the panel majority surmises that the national statistics do not, "without more, make it plausible that an African-American web developer with the educational and technical qualifications to work at NTT is more likely to have been convicted of a crime than his Caucasian counterpart," id. at 212, when it is perfectly plausible, in light of the statistics discussed above (and, indeed, the experience of the two plaintiffs in this case), that a college-educated African American applying for a position at NTT is more likely to have been convicted of a crime than his college-educated Caucasian counterpart. And finally, instead of drawing the reasonable inferences in favor of plaintiffs, the panel majority draws inferences against them, and chooses the inferences it prefers. See Menaker v. Hofstra Univ., 935 F.3d 20, 31 n.38 (2d Cir. 2019) (plaintiff must show discrimination is "one of several possible inferences," not "the most plausible" inference (internal quotation marks omitted)). 26 In short, the panel majority holds plaintiffs to a heightened pleading standard, requiring them to provide statistics relating to a specific applicant pool, even though a specific applicant pool has not yet been defined and plaintiffs have not yet had access to NTT's records. In concluding that plaintiffs failed to state a plausible claim of disparate impact discrimination, the panel majority rejects national statistics that clearly are a logical starting point for the analysis, because of its assumption that education is a "confounding variable" and its surmise that racial disparities in arrest and conviction rates dissipate as education increases. At the end of the day, after discovery and an opportunity to be heard, plaintiffs' claims may ultimately fail. But in light of the national statistics as well as plaintiffs' allegations with respect to their specific circumstances and their allegations with respect to NTT generally, their claim -- that NTT's blanket policy of denying employment to job applicants based solely on the fact of a prior conviction, without considering individual circumstances, has a disparate impact on African Americans -- is anything but implausible. 27 IV. Conclusion Our Court has long recognized the importance of Title VII and the rights it protects. 11 By denying the petition for rehearing en banc, the Court is now leaving in place a decision that will become, I am afraid, a dangerous precedent, as it will permit courts to dismiss what may be meritorious disparate impact civil rights claims because plaintiffs, who may be working "from an informational disadvantage," 975 F.3d at 212, are not able to provide "granular data" that is only in the possession of the employer, id., for an applicant pool that has not yet been defined, based on speculation that general statistics are irrelevant because the plaintiffs do not account for all purportedly "confounding variable[s]," id. at 211, all at the pleadings stage of a lawsuit. We should rehear this case en banc, vacate the judgment, and remand for plaintiffs to proceed with their claims. See Matthew Diller & Alexander A. Reinert, The Second Circuit and Social Justice, 85 Fordham L. Rev. 73, 110 (2016) (reviewing Second Circuit's social justice case law and observing that "the deeper significance of the Second Circuit's jurisprudence is its nuance and sensitivity to fundamental power inequality . . . . In 1951, Chief Judge Learned Hand captured the Second Circuit's tradition with brevity and eloquence: 'If we are to keep our democracy, there must be one commandment: Thou shall not ration justice.'"). 11 28 RAYMOND J. LOHIER, JR., Circuit Judge, joined by ROSEMARY S. POOLER, ROBERT A. KATZMANN, * DENNY CHIN, and SUSAN L. CARNEY, Circuit Judges, dissenting from the order denying rehearing in banc: Although I have very rarely voted to proceed in banc, for the reasons expressed by Judge Chin and Judge Pooler in their opinions dissenting from the denial of rehearing in banc, I dissent. The majority suggests that this case is about no more than applying the plausibility pleading standard set by Iqbal and Twombly to disparate impact cases under Title VII, and that substantive rights under Title VII itself remain unaffected. But a pleading is the key to the courthouse door. When, as here, the standard for pleadings is raised arbitrarily high or subjected almost entirely to the uneven vagaries of judicial “experience” and “common sense,” then those substantive rights that reside just behind the door wither and die. Mark the panel majority’s decision as one that will need to be revisited and corrected by us, or by Congress, in the future. Judge Katzmann concurs in Judge Lohier’s view that: “When, as here, the standard for pleadings is raised arbitrarily high or subjected almost entirely to the uneven vagaries of judicial ‘experience’ and ‘common sense,’” corrective action by the courts or Congress is in order. * 1