ASSESSING AFFIRMATIVE ACTION’S DIVERSITY RATIONALE

ASSESSING AFFIRMATIVE ACTION’S DIVERSITY RATIONALE

Ever since Justice Lewis Powell’s opinion in Regents of the University of California v. Bakke made diversity in higher education a constitutionally acceptable rationale for affirmative action programs, the diversity rationale has received vehement criticism from across the ideological spectrum. Critics on the right argue that diversity efforts lead to “less meritorious” applicants being selected. Critics on the left charge that diversity is mere “subterfuge.” On the diversity rationale’s legitimacy, then, there is precious little diversity of thought. In particular, prominent scholars and jurists have cast doubt on the diversity rationale’s empirical foundations, claiming that it rests on an implausible and unsupported hypothesis.

To assess the diversity rationale, we conduct an empirical study of student-run law reviews. Over the past several decades, many leading law reviews have implemented diversity policies for selecting editors. We investigate whether citations to articles that a law review publishes change after it adopts a diversity policy. Using a dataset of nearly 13,000 articles published over a sixty-year period, we find that law reviews that adopt diversity policies see median citations to their volumes increase by roughly 23% in the ensuing five years. In addition to exploring the effect of diversity policies on median citations, we also explore the effect of diversity policies on mean citations. When doing so, our estimates are consistently positive, but they are largely not statistically significant at conventional levels.

These findings have widespread implications. If diverse groups of student editors perform better than nondiverse groups, it lends credibility to the idea that diverse student bodies, faculties, and groups of employees generally perform better. We thus view these results as empirically supporting the much-derided diversity rationale—support that could prove critical as affirmative action confronts numerous threats.

The full text of this Article can be found by clicking the PDF link to the left.

Introduction

Debate about the relative quality of various law review volumes is usu­ally confined to student editors’ offices. On rare occasions, such discussion may spill over into the faculty lounge. But, in 2008, as then-Senator Barack Obama began his bid for the White House, the comparative merits of legal scholarship improbably became the subject of national news. Politico pub­lished an extensive article noting that Obama’s presidency of the Harvard Law Review “has generated a . . . dust-up in the blogosphere.” 1 Jeffrey Ressner & Ben Smith, Obama Kept Law Review Balanced, Politico (June 23, 2008), https://www.politico.com/story/2008/06/obama-kept-law-review-balanced-011257 [https://perma.cc/G3XY-XSCM]. In 1990, Obama became the first Black person elected to lead the Harvard Law Review in its century-long history. 2 Id. That event generated a tremendous amount of celebratory coverage at the time. 3 See, e.g., Fox Butterfield, First Black Elected to Head Harvard’s Law Review, N.Y. Times (Feb. 6, 1990), https://www.nytimes.com/1990/02/06/us/first-black-elected-to-head-harvard-s-law-review.html (on file with the Columbia Law Review); Tammerlin Drummond, Barack Obama’s Law: Personality: Harvard Law Review’s First Black President Plans a Life of Public Service. His Multicultural Background Gives Him Unique Perspective., L.A. Times (Mar. 12, 1990), https://www.latimes.com/archives/la-xpm-1990-03-12-vw-74-story.html (on file with the Columbia Law Review). But eighteen years later, some observers suggested that Obama had presided over a notoriously weak vol­ume of legal scholarship. One commenter on a legal blog counted the total number of citations to Obama’s volume and to the adjacent volumes and concluded that “Obama’s [Volume] 104 is the  least-cited  volume  of the  Harvard  Law  Review  in the last 20 years.” 4 Ressner & Smith, supra note 1 (quoting LawStatMan, Comment to Barack Obama and the Harvard Law Review, Volokh Conspiracy (Feb. 6, 2008), https://web.archive.org/​web/20080207110749/http://volokh.com/posts/1202117776.shtml (on file with the Columbia Law Review)). Another commenter went further, stating that Obama “presided over a general ‘dumbing down’” of the Review’s standards. 5 Id. (quoting Ferry Pellwock, Comment to Barack Obama and the Harvard Law Review, Volokh Conspiracy (Feb. 5, 2008), https://web.archive.org/web/​20080207110749/​http:/volokh.com/posts/1202117776.shtml (on file with the Columbia Law Review)).

Though it went unstated, the subtext of these remarks was clear: In selecting a Black student as president of the Harvard Law Review, the edi­tors who ran that journal had sacrificed quality at the altar of diversity. The implication was that Obama’s volume garnered low citations not just by random chance, or because his skills lay more with politics than with se­lecting law review articles, but because he had been awarded the position in part due to his race. There was, of course, zero evidence that race aided Obama’s membership on the law review or his ascension to its presidency. 6 Here is Obama on the question in 2000:
“I have no way of knowing whether I was a beneficiary of affirmative action either in my admission to Harvard or my initial election to the Review . . . . If I was, then I certainly am not ashamed of the fact, for I would argue that affirmative action is important precisely because those who benefit typically rise to the challenge when given an opportunity. Per­sons outside Harvard may have perceived my election to the presidency of the Review as a consequence of affirmative action, since they did not know me personally.”
The First Black President of the Harvard Law Review, J. Blacks Higher Educ., Winter 2000–2001, at 22, 24 [hereinafter JBHE, First Black President].
But the idea that there might be a tradeoff between “merit” and diversity was by then so well-ingrained into the public consciousness that the accu­sation had force even without evidence. 7 See, e.g., Fisher v. Univ. of Tex. at Austin (Fisher I), 570 U.S. 297, 331 (2013) (Thomas, J., concurring) (“Blacks and Hispanics admitted to the University as a result of racial discrimination are, on average, far less prepared than their white and Asian class­mates.”); Grutter v. Bollinger, 539 U.S. 306, 371–72 (2003) (Thomas, J., concurring in part and dissenting in part) (“The Law School is not looking for those students who, despite a lower LSAT score or undergraduate grade point average, will succeed in the study of law. The Law School seeks only a facade—it is sufficient that the class looks right, even if it does not perform right.”); Transcript of Oral Argument at 51–52, Grutter, 539 U.S. 306 (No. 02-241) (statement of Scalia, J.) (“The people you want to talk to are the high school seniors who have seen . . . people visibly less qualified than they are get into prestigious institutions where they are rejected.”); Melvin I. Urofsky, The Affirmative Action Puzzle: A Living History From Reconstruction to Today 467 (2020) (“Allowing less qualified people into college or professional school or jobs demeans the institutions themselves. How much faith can we have in the graduates of a medical school when we learn that its affirmative action admits had considerably lower qual­ifications than the norm?”).

A decade later, similar accusations resurfaced, this time in the form of a lawsuit. In early October 2018, groups of students, faculty, and alumni filed lawsuits against the Harvard Law Review and the New York University Law Review opposing practices that, they claimed, had “subordinated aca­demic merit to diversity considerations.” 8 Complaint at 6, Fac., Alumni & Students Opposed to Racial Preferences v. N.Y. Univ. L. Rev., No. 1:18-cv-9184 (S.D.N.Y. Oct. 7, 2018), 2018 WL 4899065 [hereinafter Complaint Against NYULR]; Complaint at 6, Fac., Alumni & Students Opposed to Racial Preferences v. Harvard L. Rev., No. 1:18-cv-12105 (D. Mass. Oct. 6, 2018), 2018 WL 5148474 [hereinafter Complaint Against HLR]. The lawsuits claimed that the law reviews’ policies to diversify their editorial boards have reduced the quality of the articles that the law reviews published by ensuring that articles “are judged by less capable students” and by “dilut[ing] the quality of the stu­dents who edit an author’s” work. 9 Complaint Against NYULR, supra note 8, at 6; Complaint Against HLR, supra note 8, at 6. The lawsuits also make other claims, including that the law reviews give “preference to articles written by women and racial minorities.” Complaint Against NYULR, supra note 8, at 6; Complaint Against HLR, supra note 8, at 6. This Article focuses on policies aimed at increasing the diversity of the law review editors because it is that policy that more directly implicates the broader issue of diversity in higher education. For further discussion of the lawsuits, see Bob Van Voris, Harvard Law Review Suit Opens New Front in Admissions-Bias Fight, Bloomberg (Oct. 8, 2018), https://www.bloomberg.com/news/articles/2018-10-08/​harvard-law-review-suit-opens-new-front-in-admissions-bias-fight (on file with the Columbia Law Review).

These allegations implicate longstanding criticisms of diversity initia­tives. Ever since Justice Lewis Powell’s 1978 opinion in Regents of the University of California v. Bakke made diversity in higher education a consti­tutionally acceptable rationale for affirmative action programs, 10 438 U.S. 265, 311–15 (1978) (opinion of Powell, J.). critics of affirmative action have pitted diversity against ideals of merit. They have argued that efforts to attain diversity will necessarily lead to lower quality results, as “less meritorious” individuals are selected in place of people with ostensibly stronger qualifications. 11 See infra section I.B.

Moreover, scholars and jurists who are critical of affirmative action have frequently cast doubt on the diversity rationale’s empirical founda­tions. In 2014, Professor Peter H. Schuck of Yale Law School contended: “[T]he premises underlying the diversity rationale for race-based affirma­tive action are empirically tenuous and theoretically implausible.” 12 Peter H. Schuck, Assessing Affirmative Action, 20 Nat’l Affs. 76, 76 (2014). That same year, scholar Abigail Thernstrom similarly asserted that “the entire edifice of [affirmative action] is built on a purely speculative promise that ‘diversity’ will bring educational benefits.” 13 Abigail Thernstrom, Questioning the Rationale for Affirmative Action, 16 Virtual Mentor 495, 495 (2014). Two years later, noted econo­mist Thomas Sowell voiced a particularly acerbic version of this skepticism: “Nothing so epitomizes the politically correct gullibility of our times as the magic word ‘diversity.’ The wonders of diversity are proclaimed from the media . . . and confirmed in the august chambers of the Supreme Court of the United States. But have you ever seen one speck of hard evidence to support the lofty claims?” 14 Thomas Sowell, The ‘Diversity’ Fraud, Creators (Dec. 20, 2016), https://www.​creators.com/read/thomas-sowell/12/16/the-diversity-fraud [https://perma.cc/38N8-FXXT].

Such skepticism has made its way to the august chambers of the Supreme Court as well. In 2016, Justice Samuel Alito dissented in an opin­ion upholding affirmative action at the University of Texas at Austin. In doing so, Justice Alito expressed frustration that those “invoking ‘the edu­cational ben­efits of diversity’” had “not identif[ied] any metric that would allow a court to” assess whether the purported benefits were being real­ized. 15 See Fisher v. Univ. of Tex. at Austin (Fisher II), 136 S. Ct. 2198, 2215 (2016) (Alito, J., dissenting). Even some supporters of affirmative action have doubted the qual­ity of empirical evi­dence that has been marshaled to support the diversity rationale. For ex­ample, Professor Randall Kennedy of Harvard Law School wrote a book defending affirmative action, titled For Discrimination, in which he con­fessed: “I remain doubtful about social scientific ‘proof’ of diversity’s val­ues; much of that [research] seems exaggerated and pre-determined with litigation in mind.” 16 Randall Kennedy, For Discrimination: Race, Affirmative Action, and the Law 103 (2013) [hereinafter Kennedy, For Discrimination].

Professor Kennedy is correct that the existing empirical literature on diversity in higher education is lacking. A major reason is the difficulty of measuring and evaluating performance in a manner that would shed light on the value of diversity in higher education. How can one measure the output of a university (or a law school)? Suppose a law school is able to attract a particularly diverse class for a given year. How would we be able to tell if the class’s diversity improved the students’ educational experi­ence? Given that students are graded on a curve relative to each other, assessing whether having a more diverse class improves outcomes is not a straightforward task. Accordingly, the empirical evidence on the effects of diversity has remained lacking, and the debate has raged on. 17 For a survey of the existing literature, see infra section I.C.

It is difficult to exaggerate the stakes of this debate. In January 2022, the Supreme Court agreed to hear lawsuits challenging the affirmative ac­tion admissions policies at Harvard University and the University of North Carolina. 18 See Adam Liptak & Anemona Hartocollis, Supreme Court Will Hear Challenge to Affirmative Action at Harvard and U.N.C., N.Y. Times (Jan. 24, 2022), https://www.nytimes.​com/2022/01/24/us/politics/supreme-court-affirmative-action-harvard-unc.html (on file with the Columbia Law Review). In Justice Sandra Day O’Connor’s 2003 majority opinion in Grutter v. Bollinger upholding the University of Michigan Law School’s af­firmative action policy, she suggested that affirmative action would no longer be necessary in 2028. 19 539 U.S. 306, 343 (2003) (“We expect that 25 years from now, the use of racial preferences will no longer be necessary to further the interest approved today.”). The Supreme Court in Grutter upheld the University of Michigan Law School’s affirmative action policy because it was narrowly tailored to achieve Michigan’s compelling interest in achiev­ing a diverse class. See id. at 334. By contrast, in the companion case of Gratz v. Bollinger, the Court invalidated the University of Michigan’s undergraduate admissions program because it reasoned that the program too closely resembled an impermissible quota. See 539 U.S. 244, 271–72 (2003). The Court now could very well be poised to eliminate the practice five years ahead of that schedule. 20 See Liptak & Hartocollis, supra note 18. This dispute im­plicates private institutions as much as it does public ones: The Court held in Bakke that Title VI of the Civil Rights Act, which prohibits racial discrim­ination by any private institution that accepts federal funding, is coexten­sive with the Fourteenth Amendment’s Equal Protection Clause. 21 Regents of the Univ. of Cal. v. Bakke, 438 U.S. 265, 287 (1978) (opinion of Powell, J.) (“In view of the clear legislative intent, Title VI must be held to proscribe only those racial classifications that would violate the Equal Protection Clause or the Fifth Amendment.”); id. at 352–53 (Brennan, J., concurring in part) (“Title VI’s definition of racial discrimination is absolutely coextensive with the Constitution’s . . . .”). If diversity provides no benefits, then, under current doctrine, it cannot serve as a compelling governmental interest. If it cannot serve as a com­pelling governmental interest, then affirmative action programs through­out higher education rest on an infirm foundation and may soon fall.

This Article aims to offer empirical evidence of the effects of diversity in higher education. Indeed, we believe that the lawsuits against the Harvard Law Review and New York University Law Review offer an ideal setting in which to study the influence of diversity policies on group per­formance in higher education. The work of law review editors involves core higher-education functions: Students work together in a group to evaluate the merits of scholarly work, selecting a few articles to publish from thou­sands of submissions. 22 See, e.g., Yale L.J., When to Submit Articles and Essays 2, https://​www.​yalelawjournal.​org/​files/WhenToSubmit_4xqshn68.pdf [https://perma.cc/Y9U4-HBK6] (last visited Sept. 28, 2021). The editors then edit the substance and style of those articles before publishing them. 23 See Anne Enquist, Substantive Editing Versus Technical Editing: How Law Review Editors Do Their Job, 30 Stetson L. Rev. 451, 452 (2000). The articles that student-run law reviews publish have the advantage of presenting a publicly observable out­come: arti­cle impact, as measured by citations. 24 See, e.g., Most Cited Journals and Journal Articles in HeinOnline—Updated!, HeinOnline Blog (May 4, 2009), http://heinonline.blogspot.com/2009/04/most-cited-journals-and-journal.html [https://perma.cc/X2RL-Z84D]. And although diversity policies for selecting editors may only result in a few additional diverse law students selected as editors, research suggests that even minimal increases in diver­sit­y can radically change group decisionmaking. 25 For other examples of increases in group diversity changing group behavior, see, e.g., Renée B. Adams & Daniel Ferreira, Women in the Boardroom and Their Impact on Governance and Performance, 94 J. Fin. Econ. 291, 293 (2009) (finding that “gender-diverse boards are tougher monitors”); Shamena Anwar, Patrick Bayer & Randi Hjalmarsson, A Jury of Her Peers: The Impact of the First Female Jurors on Criminal Convictions, 129 Econ. J. 603, 607 (2019) [hereinafter Anwar et al., A Jury of Her Peers] (finding that “female repre­sentation on juries significantly increased conviction rates for sex offence cases”).

We specifically investigate whether the citations to articles that a given law review publishes change after the adoption of a diversity policy for se­lecting editors. 26 These policies have been referred to in numerous ways, including affirmative ac­tion policies, racial and gender preferences, and diversity policies. We use the term “affirm­ative action” in the title because we believe it is more widely understood, but we also use the term “diversity policies” throughout this Article in part because we believe it is less politically charged. To do so, we documented the adoption of diversity poli­cies by the flagship law reviews for the top twenty law schools since 1960. We compiled a dataset of the citations to the nearly 13,000 articles pub­lished by leading law reviews during this period. In our preferred specifi­cation, we assess changes in the citations of articles published from the five years before a change in a journal’s diversity policy relative to the five years afterward. 27 We performed a robustness check in which we used different windows of time and find consistent results. See infra Part III. We find that law review membership diversity policies increase median article citations by roughly 25%. In addition to exploring the effect of diversity policies on median citations, we also explore the effect of di­versity policies on mean citations. When doing so, our estimates are con­sistently positive, but they are largely not statistically significant at conventional levels.

Notably, we find that this increase does not appear to be driven by a change in the share of articles a journal publishes on different subjects. One could imagine a diversity policy resulting in a journal publishing more articles on (for instance) constitutional law and fewer on tax law. Because citations systematically vary between subjects—for instance, con­stitutional law articles are  cited  more  than  tax  law  articles 28 See Ian Ayres & Fredrick E. Vars, Determinants of Citations to Articles in Elite Law Reviews, 29 J. Legal Stud. 427, 431 (2000). —the increase in citations to journals that adopted diversity policies could be driven by changes in the subjects of the articles being published. Nevertheless, we find no evidence that journals systematically changed the mix of subjects among the articles they accepted.

These findings have implications beyond the law review setting. If di­verse groups of student editors perform better than nondiverse groups, it lends credibility to the idea that diverse student bodies, diverse student organizations, diverse faculties, diverse teams of attorneys, and diverse teams of employees generally could perform better than nondiverse teams. We thus view these results as placing empirical heft behind Justice Powell’s much-derided rationale from Bakke. To the extent that the results are generalizable, courts should continue to view diversity as a compelling governmental interest when adjudicating affirmative action cases under the Equal Protection Clause or under Title VI.

Before continuing, it is important to acknowledge that citations are not a perfect measure of an article’s impact, much less a perfect measure of its quality. In some instances, excellent work is no doubt lowly cited and execrable work is highly cited. Nevertheless, citations are a widely used measure of research impact across different disciplines 29 See, e.g., Dennis M. Gerrity & Richard B. McKenzie, The Ranking of Southern Economics Departments: New Criterion and Further Evidence, 45 S. Econ. J. 608, 611–13 (1978) (using citations to rank economics departments); Daniel S. Hamermesh, Citations in Economics: Measurement, Uses, and Impacts, 56 J. Econ. Literature 115, 125–42 (2018) (using citations to evaluate economics articles, subfields, journals, professors, and depart­ments); John Mingers & Fang Xu, The Drivers of Citations in Management Science Journals, 205 Eur. J. Operational Rsch. 422, 422 (2010) (“Measuring the scientific impact of research­ers’ work is a difficult but important issue . . . . Particular attention has been paid to the number of citations that a publication receives.”); Gregory Sisk, Valerie Aggerbeck, Nick Farris, Megan McNevin & Maria Pitner, Scholarly Impact of Law School Faculties in 2015: Updating the Leiter Score Ranking for the Top Third, 12 Univ. St. Thomas L.J. 100, 100 (2015) (“[T]he ‘Scholarly Impact Score’ for a law faculty is calculated from the mean and the median of total law journal citations over the past five years to the work of tenured mem­bers of that law faculty.”); Iman Tahamtan, Askar Safipour Afshar & Khadijeh Ahamdzadeh, Factors Affecting Number of Citations: A Comprehensive Review of the Literature, 107 Scientometrics 1195, 1196 (2016) (“When a particular paper is cited more frequently than others, it is usually concluded that it has a higher quality compared to other papers.” (cita­tion omitted)). We further discuss the limitations and shortcomings of citations as a measure of output. See infra notes 200–202 and accompanying text. and a reasonable proxy for academic impact, as they indicate the extent to which an article has shaped the scholarly conversation. 30 See, e.g., Gerrity & McKenzie, supra note 29, at 610 (“[T]he total number of cita­tions a person or department has accumulated over a period of time is . . . a reasonably good proxy measure of the productivity of individual economists or departments.”); Mark J. McCabe & Christopher M. Snyder, Does Online Availability Increase Citations? Theory and Evidence From a Panel of Economics and Business Journals, 97 Rev. Econ. & Stat. 144, 144 (2015) (“Understanding the market for academic journals is important to scholars because it is the one market in which they function as both producers and consumers. Citations are the currency in this market . . . .” (footnote omitted)). In addition, citation counts are particularly pertinent because law review editors themselves consider how likely a piece may be cited when deciding whether to publish it. 31 See Jason P. Nance & Dylan J. Steinberg, The Law Review Article Selection Process: Results From a National Study, 71 Alb. L. Rev. 565, 585 (2008) (“Editors have an incentive to publish not the ‘best’ scholarship, but that which will be most widely read and cited.”). Thus, the measure used here is one that law review editors themselves prioritize. It is advantageous to assess the performance of law review boards accord­ing to a metric they value themselves. We also have it on good authority that citations are important to law professors, who endeavor to place their articles in publications with high visibility in order to maximize their cita­tions. 32 See James Lindgren & Daniel Seltzer, The Most Prolific Law Professors and Faculties, 71 Chi.-Kent L. Rev. 781, 783 (1996) (noting before unveiling a list of the most cited law reviews that “many law professors will use the lists of law reviews to help them determine which law reviews to submit articles to”); Joe Palazzolo, The Most-Cited Law Review Articles of All Time, Wall St. J. (June 1, 2012), https://www.wsj.com/articles/BL-LB-42728 (on file with the Columbia Law Review) (“The law professor equivalent of career hits is the ‘number of times cited’ in journals.”); Paul Caron, How to Juice Your Citations in the HeinOnline/U.S. News Rankings, TaxProf Blog (Mar. 29, 2021), https://taxprof.typepad. ​com/​taxprof_blog/2021/03/how-to-juice-your-citations-in-the-heinonlineus-news-rankings.​html [https://perma.cc/P63S-N3D6] (recommending that au­thors aim to publish in top journals in order to “juice” their citations). Unsurprisingly, this phenomenon is not con­fined to legal academia. See Tahamtan et al., supra note 29, at 1196 (“Researchers publish their findings so that they can attract the greatest attention and have the highest impact on the scientific community. They often try to publish their papers in high-impact journals to reach more readers and to become more frequently cited.” (citation omitted)).

This Article proceeds as follows. Part I sketches the legal history of the diversity rationale for affirmative action programs in the United States, which predates Justice Powell’s opinion in Bakke. It then lays out the many criticisms of Powell’s opinion that writers and thinkers across the political spectrum have lodged. In addition, Part I surveys the existing empirical literature on diversity. Part II describes the data collected for this study. Part III explains the research design used to assess the relationship be­tween the adoption of diversity policies and article citations. Part IV pre­sents the primary results, as well as a variety of checks on the robustness of those results. Part V connects the results to the broader debate over diver­sity and affirmative action. We argue that the law should provide even more space for institutions to pursue policies that will add to their diversity while simultaneously improving their work and results.