The Yale Law Journal

June 2014

A Revolution at War with Itself? Preserving Employment Preferences from Weber to Ricci

abstract. Two aspects of the constitutional transformation Bruce Ackerman describes in The Civil Rights Revolution were on a collision course, one whose trajectory has implications for Ackerman’s account and for his broader theory of constitutional change. Ackerman makes a compelling case that what he terms “reverse state action” (the targeting of private actors) and “government by numbers” (the use of statistics to identify and remedy violations of civil rights laws) defined the civil rights revolution. Together they “requir[ed] private actors, as well as state officials, to . . . realize the principles of constitutional equality” and allowed the federal government to “actually achieve egalitarian advances in the real world.” Within the frame of Ackerman’s study, these features of the civil rights revolution worked in tandem, perhaps even synergistically, helping to generate the period’s remarkable changes in voting, employment, and education. But as this essay shows, at least in the case of employment discrimination, reverse state action quickly became a threat to government by numbers. In the 1970s, no sooner did numerical measures take hold in preventing, settling, and remedying employment discrimination than courts faced claims that these measures violated the very laws pursuant to which they had been adopted. Exactly when and where state action adhered ultimately helped decide the viability of the numerical approach in the new employment discrimination regime. The eventual tensions between the “government by numbers” and “reverse state action” strands of Ackerman’s account raise questions about the content and viability of the civil rights revolution he documents. They also underscore the importance of refining his theory’s account of what he terms consolidation, synthesis, and judicial betrayal.

author. Assistant Professor of Law and History, University of Pennsylvania. I would like to thank Maria da Silva for outstanding research assistance; Christine Jolls and Reva Siegel for helpful comments; and the Journal’s editors for their help polishing the piece. I am deeply indebted to Bruce Ackerman for inviting me to participate in this symposium and for being such an inspirational teacher and scholar.


In January 1973, the Equal Employment Opportunity Commission (EEOC) reached a pathbreaking settlement with American Telephone & Telegraph (AT&T). The agency, created to implement the employment discrimination title (Title VII) of the 1964 Civil Rights Act,1 had a rough start, plagued by uneven leadership, insufficient staff, and underfunding that resulted in enormous backlogs.2 The AT&T settlement marked a turn in its fortunes. The case was sui generis. As the district court asked to approve the agreement noted, “there was no significant pending litigation in the federal courts when this consent decree was signed or . . . when the extensive negotiations that led to the decree took place.”3 The EEOC’s Chairman, William H. Brown III, called it “the most significant legal settlement in the civil rights employment history.”4 The agency’s “comparatively low profile . . . was strikingly altered” by the agreement.5 Nowhere was this more the case than within the business community: “There is a lot of teeth-chattering going on around here,” a vice-president of a large retail chain observed; equal employment consulting firms multiplied as their market of worried employers grew.6

The AT&T case plays a small but pivotal role in Bruce Ackerman’s compelling new book, The Civil Rights Revolution.7 This is the third volume in his series contending that Americans have developed an alternate system of “higher-lawmaking” in response to the near impossibility of formally amending the Constitution via Article V. According to Ackerman, all three branches periodically interact with each other and a mobilized electorate in a sustained way over time to formulate new extra-textual constitutional commitments.8 His latest installment claims that the “Second Reconstruction” was an instance of this extra-Article V amendment process and elaborates the substance of the resulting commitments.9 One of Ackerman’s key claims is that rather than adopt a universal, abstract notion of equality, Americans employed what he calls a “sphere-by-sphere” approach in which the legitimate means for achieving equality were fitted to different structures of inequality. So, for instance, although there was a general commitment to achieving “real-world egalitarian gains,”10 Ackerman explains that the constitutionally acceptable means for doing so varied for voting rights as opposed to public accommodations.11

For Ackerman, the AT&T case typified one foundational feature of the constitutional revolution regarding employment discrimination and helped forge another. First, by targeting private employers such as AT&T, Ackerman argues that Title VII fundamentally altered “the state action doctrine of the nineteenth century,” which he notes typically “insulate[d] private actors from [the Constitution’s] egalitarian principles but impose[d] them rigorously on all state actors.”12 Second, the AT&T case helped refine how the employment discrimination dimension of this constitutional revolution would employ what Ackerman refers to as “government by numbers”: the use of statistics to identify and remedy civil rights violations.13 In a 20,000-page report replete with statistical analyses, the EEOC used AT&T to model employment discrimination systemically and nationwide.14 The AT&T settlement also set numerical hiring goals and timetables for their fulfillment.15 According to Ackerman, the acceptance of the AT&T case in the executive branch and Congress secured the legitimacy of using numerical measures to remedy, not only identify and prove, employment discrimination.16

Ackerman is right to highlight the significance of the AT&T settlement. As described above, it was the kind of lightning-rod case that could generate the public and governmental response that fuels Ackerman’s theory. Going forward, it also became the template for government enforcement of Title VII.17 During the summer of 1973, the EEOC filed 150 cases against major corporations and assembled a national team of lawyers to “do the same thing as [AT&T] all over again.”18 As Ackerman observes, over the next few years, the EEOC secured court-approved consent decrees implementing similar number-based remedies in a range of industries.19

The AT&T case, however, also illuminates unexplored tensions within Ackerman’s argument. In the two years after the federal district court approved the AT&T agreement, a number of the company’s unions petitioned the court to modify its consent decree.20 They contended, among other things, that by imposing numerical goals, the consent decree violated Title VII and the Constitution’s equal protection guarantees.21 The district court easily dismissed their claims in 1976, citing abundant precedent.22 The Third Circuit Court of Appeals quickly affirmed, but cautioned that preferential remedies were not categorically constitutional and “must be held invalid under the . . . Fifth Amendment unless” they satisfied “strict scrutiny.”23 In the appeals court’s judgment, the AT&T agreement met this test.24 The Supreme Court had not yet applied strict scrutiny to racial preferences adopted to remedy or prevent discrimination.25 But over the next decade, the Supreme Court found an increasing number of preferential schemes to be unconstitutional.26 During this same period, the Court gave greater leeway to preferences under Title VII.27

The “reverse-state-action” and “government-by-numbers” aspects of Ackerman’s constitutional revolution were on a collision course. With preferential treatment more liberally allowed under Title VII than under the Constitution, using numbers to address employment discrimination became incompatible with applying the Constitution to the private sector. Instead, ensuring that the state action doctrine “insulate[d] ‘private’ actors from the [Constitution’s] egalitarian principles” ended up protecting the “government-by-numbers” approach to employment discrimination.28

Part I explains in greater depth Ackerman’s claim that the civil rights revolution included a major reworking of the state action doctrine and an embrace of numerical approaches to achieving equality. It also explains why employment is a particularly ripe site to explore the tensions between them. Part II demonstrates how the Court’s preservation of the nineteenth-century state action doctrine helped preserve the use of numbers to achieve equality in the workplace when employers’ voluntary use of preferences first came before the Court in the 1970s. Part III explains how this state action shield further protected these preferences in the 1980s, both before the Supreme Court and within President Ronald Reagan’s administration. Part IV argues that the state action doctrine could once again rescue voluntary employer preferences. The Conclusion raises some possible implications of this history for the stability of the civil rights revolution Ackerman documents and for his broader theory of constitutional change.

I. reverse state action and government by numbers in ackerman’s revolution

Ackerman’s civil rights revolution has three key characteristics: it took a “sphere-by-sphere” approach to codifying the Constitution’s equality guarantee, it employed “government by numbers” to achieve this constitutional goal (albeit to varying degrees in each of the spheres), and it extended this constitutional mandate to the private sector. The result was targeted statutes and tailored administrative regimes customized to the particular structures of inequality in voting, public education, housing, public accommodations, and employment. As I explain below, these latter two characteristics coincided to the greatest degree in employment, making it a prime site to explore the tensions between them.

A. Sphere by Sphere

Ackerman argues that the Supreme Court’s increasingly consistent insistence that equal protection is achieved through colorblind policies betrays a central tenet of the civil rights revolution.29 Ackerman traces this tenet back to Chief Justice Earl Warren’s opinion in Brown v. Board of Education.30 “In making its case against ‘separate-but-equal,’” Ackerman contends, Brown’s reasoning “doesn’t apply to all social relationships across the board” but instead “requires the law to single out crucial spheres for the vindication of equality.”31 When Congress enacted a series of civil rights statutes in the 1960s and early 1970s, it adopted this approach, Ackerman argues, “self-consciously divid[ing] the world into different spheres of life: public accommodations, education, employment, housing, voting.”32 Congress’s goal in each sphere was the same: “the pursuit of real equality of opportunity.”33 But Congress had a “contextual understanding of the constitutional meaning of equality in different spheres of social and political life.”34 As a result, it “displayed great creativity in crafting different administrative setups for different spheres,” which in turn “led to the development of different rules and principles in different spheres.”35 For Ackerman, it was these “landmark statutes,” with their sphere-by-sphere approach, not the Court’s increasingly formalist, homogenous approach to equality in decisions such as Loving v. Virginia, that defined the civil rights revolution and merit “full constitutional status.”36

B. Government by Numbers

Congress, the executive branch, and the courts deployed “government by numbers” to varying degrees in each sphere of discrimination they targeted. One of Ackerman’s great insights is how the civil rights revolution brought the technocratic impulse of the New Deal to bear on the problem of racial inequality.37 Government by numbers was used, Ackerman argues, only where necessary “to bridge the yawning gap between the law on the books and the law of ordinary life.”38 Generally, this involved discrimination by “institutions whose internal logics of decision are opaque”39 or who were structurally disinclined to comply.40 Government by numbers was used paradigmatically in the Voting Rights Act, which “imposed hard-edged output tests on voting registrars” and extensive federal supervision on those that failed.41 In contrast, it played no role in policing public accommodations, where discrimination was patent and the market rewarded compliance.42 Here, Congress “principally relied on traditional lawsuits by traditional plaintiffs before traditional courts using traditional legal language to attack discrimination.”43

But government by numbers was not an all or nothing proposition, Ackerman reveals. True to the sphere-by-sphere approach, in the case of employment discrimination, the civil rights revolution settled on an intermediate position. As initially enacted, Title VII focused on individual intentional discrimination and forbade numerical quotas, “the hard-edged version of government by numbers” used in voting.44 But Congress did not banish numbers entirely. They could be—and were—used as evidence of discriminatory intent,45 to identify suspicious employer policies, and to target the EEOC’s limited enforcement resources on the “worst offenders.”46 The EEOC also used them to demonstrate that discrimination was a systemic, not individual or intentional, problem.47 These uses in turn encouraged businesses to tend to their numbers and to adopt policies that lowered any numerical red flags. “Within the setting of the modern enterprise,” Ackerman observes, “discriminatory intentions had become a numbers-driven affair.”48 First the Supreme Court in Griggs v. Duke Power,49 then Congress and President Richard Nixon endorsed these uses of numbers in the 1972 amendments to Title VII.50 The result was a cascade of firms seeking consent decrees, adopting affirmative action programs, and being subject to judgments that used numbers to create preferences for hiring, training, and promoting minorities and women.51

C. Reverse State Action

Another central claim of Ackerman’s is that the civil rights revolution cast aside the state action limit on the Constitution’s equality guarantee. Ackerman argues that the Supreme Court “was on the verge of dispatching the state action doctrine in its 1964 decision in Bell v. Maryland,” a case involving the trespassing convictions of sit-in protesters at a Baltimore, Maryland restaurant.52 But he acknowledges that it “refrained at the last minute since Congress was poised to take leadership on this issue with the Civil Rights Act.”53 Nor did the Court in his account later take the step it had avoided in Bell. Nonetheless, he contends that the “state action restriction simply doesn’t make sense of the living Constitution.”54 For Ackerman, the Constitution includes the commitments “expressed in the landmark [civil rights] statutes.”55 And these “requir[ed] private actors as well as state officials to accept wide-ranging responsibilities to realize the principles of constitutional equality.”56 Indeed, from Title VII, to the Civil Rights Act’s public accommodations title, to the Fair Housing Act, these statutes engaged in what Ackerman calls “reverse state action,” singling out the private sector.57 Ackerman notes that “[t]he current legal community uses Bell’s deference to Congress as an excuse for pretending that the state action doctrine of the nineteenth century remains intact.”58 He urges, however, that doing so is “[a] big mistake.”59


Neither reverse state action nor government by numbers characterizes every sphere Ackerman analyzes. As explained above, reverse state action pertained only to employment, housing, and public accommodations; of these, government by numbers was employed most thoroughly in the employment discrimination context.60 Within the timeframe of Ackerman’s account, these two aspects of the civil rights revolution coexisted harmoniously. But no more than a few years thereafter, harmony gave way to conflict.

II. the state action doctrine preserves government by numbers in weber

When the Third Circuit applied strict scrutiny to the AT&T consent decree in 1977, it was stepping out ahead of the Supreme Court, which had yet to adopt this standard of review for racial preferences designed to remedy racial discrimination. But the next year, in Regents of the University of California v. Bakke, the Court signaled that it was considering the Third Circuit’s approach and sounded a warning about remedial use of preferences under the landmark civil rights statutes.61 As this Part recounts, Bakke was but the opening salvo in a complicated, multi-front reconsideration of government by numbers in the employment context. Over the next decade, the government-by-numbers dimension of Title VII clashed repeatedly with the Supreme Court’s increasingly colorblind interpretations of the Constitution. When it came to employers’ voluntary use of preferences, the Court’s formalist, narrow approach to state action helped preserve the government-by-numbers approach.

A. The Growing Constitutional Threat to Racial Preferences

Racial preferences in hiring, promotion, and training were already controversial in 1973 when the EEOC entered into its pathbreaking agreement with AT&T. In 1972, both presidential candidates denounced the use of quotas and even some prominent liberals criticized the Nixon administration’s use of hiring goals.62 The Nixon Administration described the former as a number that “managers are obligated to attain” and the latter as one that would “probably be met.”63 To liberal opponents of quotas, however, this was a distinction without a difference. They thought that goals often functioned as “de facto quotas.” Even more fundamentally, however, they opposed any departure from meritocratic hiring. From this baseline, probable goals were no better than mandatory quotas.64 Labor leaders also criticized hiring goals, let alone quotas, for harmfully pitting white and black workers against each other. Guaranteeing jobs for all, not quotas, was the labor movement’s answer to African Americans’ limited job opportunities.65 Conservatives formulated a different critique, embracing colorblindness as the sine qua non of equal protection. Under this by now familiar formulation, recognizing race to remedy or prevent discrimination was constitutionally equivalent to recognizing race in order to discriminate.66

Bakke signaled the rising influence of the colorblind approach. In that case, the Court was asked to determine whether a university program setting a minimum quota for minority admissions violated the Equal Protection Clause and Title VI of the Civil Rights Act of 1964.67 The Justices splintered so thoroughly that it was hard to conclude anything definite from the decision other than that the admissions program was illegal. But in the Court’s first decision reviewing voluntary race-conscious measures adopted to remedy or prevent discrimination, colorblind arguments were key. Justice Powell endorsed a colorblind interpretation of equal protection, calling for strict scrutiny any time the government treated people differently based on their race.68 He found that the admissions policy failed this standard.69 Four other Justices avoided the constitutional question by interpreting Title VI in equally ominous, and fatal, colorblind terms.70

When the Court considered a similar quota-based training program adopted by a private sector employer and union the next term, it seemed possible that the Court would strike it down under a colorblind approach to Title VII, equal protection, or both. Title VII had already been interpreted in ways that defied a strictly colorblind approach, giving the program a better chance under Title VII than it would have had under Title VI.71 But this would only remove the dodge used by the Justices who struck down the admissions program in Bakke on statutory grounds. That a constitutional decision would be close was near certain: Bakke had been a five-four decision.72 That colorblind constitutionalists would command a majority striking the training program down was entirely possible. Or at least, it would have been if the Supreme Court had recognized that the civil rights revolution had swept away the state action doctrine.

B. Employer Preferences’ State Action Shield

1. Narrowing State Action

Kaiser Aluminum’s quota-based training program was saved from the constitutional gauntlet because the Supreme Court had not rejected the state action doctrine in or since Bell.73 The Court began broadening the definition of state action in the 1940s,74 and came close to vastly expanding it in Bell.75 Over the next several years, the Court continued to inch its way towards a more expansive understanding of state action,76 leading one court watcher to predict that the “radical changes in society . . . even now are tolling the demise of state action.”77 But this progression of decisions can overstate the doctrine’s vulnerability. Even liberals at the time were deeply divided over its fate.78 The Justices who signed on to the Court’s most expansive decisions insisted that they were preserving an intelligible line between public, constitutionally accountable actions and those that were private and thus free from constitutional scrutiny.79 When President Nixon replaced Earl Warren with a Chief Justice strongly opposed to state action’s expansive trend, he tipped the balance firmly in favor of contraction. Over the course of the 1970s, the Court issued a series of decisions narrowing the circumstances under which state action would be found,80 the most recent of which was decided the same term as Bakke.81

2. Dodging the Equal Protection Bullet in Weber

This narrowed state action doctrine allowed the Court to avoid deciding the constitutionality of Kaiser’s training program. Writing for the majority in United Steelworkers of America v. Weber, Justice Brennan began by noting that “[s]ince the Kaiser-USWA plan does not involve state action, this case does not present an alleged violation of the Equal Protection Clause.”82 He instead considered whether it furthered Title VII’s overarching purpose of opening jobs to African Americans. The state action divide otherwise critically informed Brennan’s analysis. In order to distance Title VII from Bakke’s colorblind interpretation of Title VI, Brennan insisted that Title VII as originally enacted “regulate[d] purely private decisionmaking and was not intended to incorporate and particularize the commands of the Fifth and Fourteenth Amendments.”83 When divining Title VII’s purpose, Brennan emphasized that Congress had sought to “avoid undue federal regulation of private businesses,” most notably by not “limit[ing] traditional business freedom to such a degree as to prohibit all voluntary, race-conscious affirmative action.”84 Justice Rehnquist accused Brennan of a “Houdini”-like opinion that “eludes clear statutory language [prohibiting racial discrimination], uncontradicted legislative history, and uniform precedent.”85 But a majority of the Court found that Title VII permitted the training program.86

3. A Weber What If?

By preventing the Court from considering the constitutionality of Kaiser’s program, the state action doctrine likely not only saved the program, but also either prevented a colorblind interpretation of Title VII or forestalled a strict scrutiny approach to affirmative action. The state action doctrine also channeled the Court in Weber towards an affirmative-action tolerant interpretation of Title VII that immunized voluntary employer preferences when they came under sustained attack in the 1980s. My argument relies entirely on counterfactual, which as a historian I am trained to distrust. But I am setting aside those compunctions and making a more lawyerly case.

The first (if unlikely) threat was that the Court would avoid the constitutional question by adopting a colorblind interpretation of Title VII, as the concurring Justices had for Title VI in Bakke. Justices Powell and Stevens did not participate in Weber. Brennan wrote for the majority. Of the remaining six Justices, four (Thurgood Marshall, Potter Stewart, Byron White, and Harry Blackmun) joined Brennan’s interpretation of Title VII. Two of them would thus have had to defect on the statutory interpretation to avoid having to address the constitutional question. Only one, Stewart, had joined the colorblind interpretation of Title VI in Bakke.87 For reasons discussed further below, White might have provided the second vote, securing a colorblind interpretation of Title VII.

If the constitutional question had been reached rather than avoided, it was quite possible (if far from certain) that the Court would have done what it had refrained from doing in Bakke: hold that preferences triggered strict scrutiny. This is where my counterfactual gets the most questionable from a historian’s perspective, because it relies not on predicting the future based on the past, but the past based on the future.88 But here goes. Although Weber did not press the constitutional question, Bakke was just the first in a steady stream of cases in which the Court considered the constitutionality of various affirmative action programs. In fact, one year after Weber, the Court considered a minority set-aside for federal contracts in Fullilove v. Klutznick.89 Of the seven Justices participating in Weber, two—Justices Rehnquist and Stewart—subjected the federal affirmative action program to strict scrutiny in Fullilove.90 Three, Justices Brennan, Marshall, and Blackmun, favored applying intermediate scrutiny.91 That leaves Chief Justice Burger and Justice White. Both declined to decide the proper level of scrutiny to apply in Fullilove, finding the program passed muster under either.92

There are good reasons to predict that a majority in our hypothetical Weber would favor strict scrutiny. Over the course of the 1980s, first Burger and then White signed on to the strict scrutiny standard.93 Burger’s initial openness to applying intermediate scrutiny to the federal affirmative action program in Fullilove did not mean that he would have backed a lower level of scrutiny in Weber. Instead, his opinion in Fullilove repeatedly emphasized the special deference the Court owed Congress as a co-equal branch and the uniquely broad scope of Congress’s remedial powers.94 His openness to applying intermediate scrutiny to the federal affirmative action program in Fullilove thus did not mean that he would have backed a lower level of scrutiny in Weber. Further supporting this outcome, Burger favored a colorblind interpretation of Title VI in Bakke and of Title VII in Weber.95 Justice White is a trickier case since he had favored intermediate scrutiny in Bakke. The fact that he chose to avoid the level of scrutiny question in Fullilove suggests that he was not wedded to intermediate scrutiny, however. Also, even before he signed on to strict scrutiny, he began finding that affirmative action policies denied equal protection, in agreement with the Justices that applied it.96 It is likely, then, that at least Chief Justice Burger and possibly Justice White would have provided the two remaining votes necessary to constitute a majority in favor of applying strict scrutiny in our counter-factual version of Weber. If they did, the fact that the non-participating Justice Powell was already on record supporting strict scrutiny would have made the full Court’s alignment on the issue appear secure.97 That the other non-participant, Justice Stevens, also seemed to embrace strict scrutiny the next year in Fullilove would have made the alignment starker.98

If the Court had applied strict scrutiny in Weber, it would almost definitely have found Kaiser’s program unconstitutional, demonstrating why preserving government by numbers in the workplace required creating space between the treatment of employer preferences under Title VII and equal protection. This step also involves counterfactual speculation, as the Court would settle on what standard to apply to affirmative action (strict scrutiny) before flushing out how to apply that standard.99 Applying Justice Powell’s test from Bakke and Fullilove, however, is probably a safe approach.100 As more Justices signed on to strict scrutiny during the 1980s, they complained that Powell’s version was not strict enough.101 If Weber would fail Powell’s test, therefore, it would likely fail any articulation of strict scrutiny.

Under Powell’s approach, it is unlikely that the Court would have found that the Kaiser plan furthered a compelling interest. As an end in itself, “[r]acial preference never can constitute a compelling state interest,” Powell insisted.102 Nor was the diversity rationale Powell found compelling in Bakke advanced in Weber or clearly applicable, given Powell’s heavy reliance on factors particular to education.103 Powell did allow that “ameliorating, or eliminating where feasible, the disabling effects of identified discrimination” was a compelling interest.104 But he limited it to discrimination that was identified through “judicial, legislative, or administrative findings of constitutional or statutory violations.”105 As Powell found the university admissions office in Bakke incapable of making the predicate finding, he would of necessity find Kaiser Aluminum and its unions similarly incapacitated.106 In Weber, the Court took judicial notice, based on “numerous” lower court decisions, of the craft union discrimination said to explain Kaiser’s decision to adopt its plan.107 Perhaps our hypothetical Weber court would, as the actual one did, accept prior judicial findings of discrimination in lieu of findings by the institutionally incompetent policy adopter. This still would not save the Kaiser plan, however. The actual Weber court found “break[ing] down old patterns of racial segregation and hierarchy” in “occupations which have been traditionally closed to” a protected class sufficient justification for employer preferences.108 But Powell insisted that countering this sort of “societal discrimination” could not justify a racial preference for strict scrutiny purposes.109

Even if Kaiser’s plan were somehow deemed to advance a compelling interest, it likely would have failed the narrow tailoring test. Some of the factors Weber considered to determine the plan’s comportment with Title VII were similar to those Powell employed under equal protection.110 Other factors were different but not in ways likely to lead to different outcomes.111 Powell’s narrow tailoring test imposed a unique criterion, however, which Kaiser’s plan would most certainly fail, even if it would stand under all the remaining factors112: Powell’s test required that no alternative, less “trammeling” remedies were available.113 Indeed, in the cases where Powell was willing to find that a preference, let alone a disproportionately large quota like that used in the Kaiser plan, was narrowly tailored, he did so in part because more moderate remedies had been tried and failed.114 No such history existed at Kaiser.115


The Court might have adopted a colorblind interpretation of Title VII or found that the Kaiser plan violated equal protection had the state action doctrine not foreclosed this inquiry in Weber. But this thought experiment is about more than the fate of Kaiser’s voluntary quota. The attack on affirmative action began in earnest in the 1980s when the Reagan Justice Department set out to eliminate all racial preferences. That Weber had instead read Title VII to allow the Kaiser plan provided a bulwark for employers’ voluntary preferences during the 1980s attack.

III. state action’s lingering protective effects

The actual Weber decision allowed an employer to defend against a claim of reverse discrimination by establishing that its complained of actions were part of a Title VII-compliant affirmative action program. During the 1980s, this test helped preserve voluntary affirmative action under Title VII and opened a divide between its analysis under Title VII and equal protection.

A. The Reagan Administration’s Attack on Preferences

The legal attack on preferences, thus far waged by individuals and unions, got an enormous boost when President Reagan put the resources and considerable Supreme Court capital of the executive branch behind it. Immediately after taking office, Reagan assured that his “administration is going to be dedicated to equality” but opposed to any “affirmative action programs becoming quota systems.”116 Again, after reelection, he reminded the nation that he had run against quotas in both campaigns, and claimed a mandate to eliminate them.117 As President, he appointed opponents of racial preferences to key civil rights positions, including as Chairman of the Civil Rights Commission and the EEOC.118

Reagan’s appointment of William Bradford Reynolds to head the Justice Department’s Civil Rights Division most threatened government by numbers under Title VII. Reynolds promptly critiqued “racial formulas, such as hiring quotas and fixed goals . . . in the work place.”119 These formulas “create[d] a caste system in which an individual must be unfairly disadvantaged for each person who is preferred,” Reynolds urged, and were “as offensive to standards of human decency today as they were some 84 years ago when countenanced under Plessy v. Ferguson.”120 The White House quickly assembled a special task force on employment discrimination,121 and Reynolds publicly announced his desire to, as the Wall Street Journal put it, “get the Supreme Court to rule that it is illegal and unconstitutional to give minorities and women preference in hiring and promotion.”122 The administration’s attack would be wholesale. Reynolds “bridled at government-imposed preferences and ‘race-conscious relief,’ even as a remedy for proven discrimination.”123 Instead, remedies could benefit “only the individuals discriminated against. . . . Relief, as the tag went, must be victim-specific; if so it is color-blind, for it is predicated on victim status, not race.”124 Reconsidering Weber, limiting relief under Title VII, securing strict scrutiny for preferences, and ensuring that narrow tailoring would, with exceedingly minimal exceptions, be “fatal in fact”125—all would be on the administration’s sweeping agenda.

In complicated ways, Reynolds’s plans were limited by the state action doctrine. Most obviously, the state action doctrine had secured Weber, a precedent that provided a drag on the administration’s attack over the course of the 1980s.

B. Weber’s Protective Effects

Weber delayed and sculpted the Reagan administration’s challenge to voluntary employer preferences and helped protect them once that attack occurred.

1. Equal Protection Not Title VII; Public Not Private

With Weber so recently decided, the Reagan administration initially avoided the issue of voluntary employer preferences under Title VII. When Reynolds laid out his agenda in 1981, he opined that “Weber is wrongly decided” and expressed hope that, were the Court to revisit the issue, “it would arrive at a different conclusion.”126 He also spoke out against all uses of “statistical formulae” in the workplace.127 But attacking voluntary employer preferences under Title VII would not be the administration’s immediate goal.128 For the most part, the administration challenged courts’ authority to order—not employers’ leeway to voluntarily adopt—affirmative action programs.129 To the extent that Reynolds focused on voluntary preferences, he pledged to fight for constitutional constraints on their use and focused on public, not private, employers.130 For instance, the government sought unsuccessfully to intervene before the Sixth Circuit Court of Appeals in a challenge to an affirmative action policy adopted by the Detroit Police Department.131 As the Los Angeles Times reported, the Justice Department argued “that public employers are barred by the Constitution from voluntarily setting up promotion plans based on race, even though the Supreme Court [in Weber] approved them for private employers.”132

Further, when Reynolds subsequently targeted voluntary employer preferences under Title VII during Reagan’s second term, he sought to circumvent Weber by challenging public employer programs. Reynolds’s influence grew after Reagan’s reelection, as he joined the Justice Department’s litigation steering committee and his protégés from the Civil Rights Division moved into influential positions in the Department.133 He also gained a powerful ally in Edwin Meese, a close adviser of Reagan’s who became Attorney General in 1985. Meese quickly declared that “[c]ounting by race is a form of racism.”134 The Reagan administration, he asserted, “reject[ed] . . . unequivocally” the argument that affirmative action required “race-conscious, preferential treatment.”135 Instead, using “racially preferential quotas . . . is nothing short of a legal, moral and constitutional tragedy.”136 Upon reelection, Reagan specifically singled out for attack the use of voluntary quotas under Title VII, declaring them “exactly what the civil rights laws were designed to stop.”137

Reagan’s Justice Department did not challenge Weber head on, however. Instead, it built on Justice Rehnquist’s unusual dissent when the Court in 1985 declined to review an appellate court decision upholding a state’s voluntary preferences under Weber.138 Rehnquist quoted Weber’s caveat that Kaiser’s plan “does not involve state action” and urged that the Court should address how the presence of state action affected the Weber framework.139 The preferences’ violation of equal protection was not at issue in the denied petition, Rehnquist acknowledged. But he urged that “when a state employer claims that arguably discriminatory conduct on its part is nonetheless authorized by Title VII, the claim of such statutory authorization must be considered in the light of the prohibitions of the Fourteenth Amendment.”140 Even if Congress intended private employers to have the leeway to adopt voluntary preferences that Weber provided, Rehnquist contended that it could not have intended to allow state and local employers “to claim that their actions were shielded under Title VII even if the actions would violate the Fourteenth Amendment.”141 It was this narrower point only that Reynolds pursued during Reagan’s second term.142

2. Precedential Effects

When the Justice Department eventually challenged public employers’ voluntary preferences under Title VII, Weber helped protect them. Most directly, only because of stare decisis did a majority of the Court find that the public employer preference at issue in Johnson v. Transportation Agency comported with Title VII.143 Without Justice Stevens’s joinder, Justice Brennan’s opinion would not have commanded a majority of the Court. Stevens joined despite believing that Weber was wrongly decided.144 Weber, however, was now “an important part of the fabric of our law,” Stevens explained, which he found a “sufficiently compelling” reason to “adhere to [its] basic construction of” Title VII.145 That Weber was already on the books also secured Justice O’Connor’s concurrence in the judgment, decisively ensuring the continued viability of voluntary employer preferences under Title VII.146 With the remaining three Justices all in favor of reversing Weber, its existence is all that saved the Court from adopting a colorblind interpretation of Title VII in Johnson.

Less tangibly, Weber shielded voluntary preferences through what might be called the first-mover advantage. While hardly a formal canon of construction, where other statutes regulated similar conduct but only some involved state action, the Court’s initial interpretation has bled across the state action line regardless of the different constitutional stakes.147 Imagine if the Court in Johnson had been considering for the first time voluntary employer preferences under Title VII or that my hypothetical Weber had produced a splintered Court akin to that in Bakke, with neither a constitutional nor statutory ground prevailing. If so, this first authoritative interpretation of Title VII would have occurred deeply in the Constitution’s shadow. Indeed, the government and a dissenting Justice Scalia urged the Court in Johnson to follow the approach Rehnquist had suggested previously and read equal protection limits into Title VII.148 For the reasons discussed in the Weber hypothetical, this would have led to a narrower ambit for voluntary employer preferences.149 This more crabbed interpretation, in turn, might have then been applied in the private employer context due to the first-mover effect.150 Instead, the inertial path ran in the other direction: the Court swept aside equal protection concerns in Johnson,151 and applied Weber’s test, emphatically crafted for the private sector, to the public employer involved.152


Thanks to the Court’s narrow, formal approach to state action, during the 1980s, Title VII preserved space for voluntary preferences by private employers that were no longer available to public employers under equal protection. Going forward, even the Reagan Justice Department conceded that the test under Title VII was more permissive than under equal protection.153

IV. state action to the rescue again?

In 2009 the Supreme Court in Ricci v. DeStefano did what it had refused to do in Johnson: interpret Title VII in light of equal protection standards.154 Ricci involved a reverse discrimination claim against a city for deciding not to certify the results of a promotion exam on which “white candidates had outperformed minority candidates.”155 In defense, city officials argued that, if “they had certified the results, they could have faced liability under Title VII for adopting a practice that had a disparate impact on the minority firefighters.”156 The Court avoided applying equal protection directly by finding the city’s actions violated Title VII.157 Equal protection nonetheless crept into the Court’s analysis, as Justices Rehnquist and Scalia as well as the Reagan administration had urged it should in the 1980s.158 How broadly Ricci sweeps is as yet unclear.159 As courts work this question out, the state action doctrine could once again shield some voluntary employer preferences.

Ricci newly imports equal protection standards into Title VII analysis of employers’ preferential practices. Finding that Title VII’s disparate impact and disparate treatment provisions were “in conflict absent a rule to reconcile them,”160 the Court turned to “cases similar to this one, albeit in the context of the Equal Protection Clause.”161 Specifically, Ricci borrowed these equal protection cases’ requirement that an employer have a “‘strong basis in evidence’ that [its] remedial actions were necessary” before a court will find the employer’s race-conscious efforts to remedy past racial discrimination constitutional.162 Translated into the Title VII context, in order to escape liability, Ricci required an employer to have a strong basis in evidence that its abandoned practices either “were not job related and consistent with business necessity, or” if they were, that “there existed an equally valid, less-discriminatory alternative that” the employer had “refused to adopt.”163

Thus far, Ricci has not spelled the end of voluntary employer preferences but it has created the need to distinguish the preferences to which the framework applies. Ricci does not mention Johnson or Weber,164 and the Ricci dissenters explicitly distinguish the employer defense rejected therein from that allowed in Johnson and Weber.165 Whether that distinction will hold is as yet undetermined. The majority opinion contains language that could sweep broadly to include preferences such as those in Weber and Johnson that are adopted to correct a “manifest imbalance” as well as language promising the decision’s narrow reach.166 Lower courts thus far have interpreted Ricci to apply only to public employer preferences adopted to avoid disparate impact liability.167 The Second Circuit has given the question its most extensive treatment, finding that the Weber/Johnson framework still applies if a public employer prospectively adopts preferences to benefit a protected class generally.168

As yet unaddressed is the Ricci test’s application to private employers’ preferences, whether adopted to avoid disparate impact liability or to “break down old patterns of racial segregation and hierarchy.”169 Whatever position the Supreme Court takes regarding the Second Circuit’s approach to public employers, the state action doctrine could (and indeed should) foreclose Ricci’s application to private employer preferences under either the disparate impact or Weber/Johnson defense. The strong-basis-in-evidence test that Ricci borrowed is derived from strict scrutiny’s compelling interest prong.170 Even the Court’s conservative Justices (indeed, especially the Court’s conservative Justices) should find imposing this standard on private employers unwarranted. If Ricci did no more than borrow a standard from equal protection doctrine, then as a matter of statutory construction, this borrowing is less justified in the private sector. In Ricci, the only factor the Court weighed in favor of maximizing employer flexibility was “Congress’s intent that ‘voluntary compliance’ be ‘the preferred means of achieving the objectives of Title VII.’”171 Regarding private employers, however, an additional factor weighs in favor of flexibility and against Ricci’s borrowing: what Justice Brennan in Weber described as Congress’s desire to limit federal regulation of the private sector and maximize “traditional business freedom.”172 This desire derives from the same public/private divide that animates the traditional state action doctrine.

As a matter of constitutional law, whether Ricci merely borrows equal protection doctrine or is conforming Title VII to it, the Justices in the Ricci majority should be loath to extend its test to private sector employers. If Congress intended equal protection doctrine to sculpt the scope of Title VII as applied to state and local employers has been debated since at least Johnson.173 But whatever arguments this position has in its favor, they evaporate as regards the private sector—especially for those wedded to a narrow conception of state action. State and local employers’ voluntary preferences are undeniably infected with state action and Congress, in applying Title VII to them, relied on its authority to enforce the Fourteenth Amendment’s equal protection provisions.174 In contrast, Congress grounded the Title VII provisions governing private actors solely on its Commerce Clause authority.175 Furthermore, these provisions merely permit private actors to use voluntary preferences, a form of state action even the Warren Court at its height never found to trigger constitutional scrutiny.176 There is thus no basis for importing equal protection standards into Title VII review of private employers’ voluntary preferences. Indeed, doing so would violate one of the central tenets of the state action doctrine. As Reagan’s former Solicitor General Charles Fried, a strong supporter of the state action divide and the colorblind Constitution, has argued, “the notion that governments are bound by stricter rules than are private actors” renders Weber “appropriate.”177

Even if Title VII does more than permit those preferences adopted to avoid disparate impact liability, there is still not sufficient state action to trigger constitutional concerns. Some affirmative action critics argue that Title VII not only permits, but also actually encourages employer preferences by providing disparate impact liability.178 This claim is at its strongest when an employer takes preferential action in order to avoid feared disparate impact liability, as the city in Ricci did. Yet even this strongest case should not trigger equal protection concerns in private employment, especially given the state action contractions of the Burger Court. In the late 1970s and early 1980s, Chief Justice Burger and Justice Rehnquist led the Court in more narrowly construing when a state “exercised coercive power or . . . provided such significant encouragement . . . that the choice must in law be deemed to be that of the State.”179

The Court, since articulating the “significant encouragement” test, has never found that the test was met,180 nor should proponents of a contracted state action doctrine do so in this instance. Indeed, it would vastly expand the ambit of state action were the Court to hold that a private decision made to avoid legal liability sufficed.181 Every action taken to avoid civil or criminal laws could potentially result in constitutional applicability. For instance, a private employer that prevents its employees from making racist or sexist remarks in the workplace in order to avoid hostile work environment claims would trigger free speech rights for the censured employee. If the employer bars handguns in the workplace to avoid violent workplace felonies, its employees could raise Second Amendment challenges. Just as the current Court would never find that an employee could directly bring equal protection challenges against a private employer that adopts a preference in order to avoid disparate impact liability, the Court should not import equal protection standards into Title VII analysis of those preferences.


Even if Johnson falls to Ricci, Weber should not share its fate, nor should preferences that private sector employers adopt to avoid disparate impact liability.


If and when the Supreme Court addresses Ricci’s application to private employment, proponents of voluntary employer preferences will have to decide how to limit its reach. As this history demonstrates, for over thirty years, the Supreme Court’s narrow view of state action has been these preferences’ primary protection. Proponents would be wise to insist that it remain so. I have marshaled this history not to map future litigation strategy, however, but to emphasize how quickly and persistently the reverse-state-action and government-by-numbers dimensions of Ackerman’s civil rights revolution came into conflict. With government by numbers secured under Title VII but increasingly constrained under equal protection, the relationship between the two became of paramount importance. The narrow state action doctrine fashioned by the Burger Court directly and indirectly insulated and preserved government by numbers under Title VII and in the private sector.

What does this mean for Ackerman’s account of the civil rights revolution? I began by arguing that this history reveals an unexplored tension inherent in Ackerman’s account. If Title VII truly reworked the state action status of private employers and unions, then most likely either Title VII would have been interpreted to prohibit voluntary preferences or they would have been found unconstitutional. Either way, a key dimension of government by numbers would have been eliminated from the civil rights revolution’s response to employment discrimination.182 If so, the civil rights revolution, as Ackerman describes it, was inherently unstable.

The tension could be illusory, but explaining it away raises hard and as yet unanswered questions for Ackerman’s overall theory of extra-Article V constitutional amendment. One way to explain away the tension is to contend that it resulted not from conflicts endemic to the civil rights revolution but from the Court’s failure to appropriately incorporate that revolution’s constitutional changes. In this account, the Court betrayed the civil rights revolution when it turned against government by numbers and towards colorblindness in the equal protection context.

This is a plausible, even attractive, position: scholars have lodged trenchant critiques of the Court’s colorblind turn since it began.183 But this account raises questions about how consolidation is measured under Ackerman’s theory. According to Ackerman, the Court never fully embraced government by numbers under equal protection.184 Instead, it wavered, reaching back to what Ackerman calls the anti-humiliation principle in Brown even as it deployed a colorblind constitutionalism dating to World War II.185 None of the cases Ackerman discusses involved the workplace. The Court’s first decision addressing government by numbers under equal protection and in the employment discrimination context, Washington v. Davis, rejected it.186 Coming in 1976, shortly after Ackerman’s account ends, Davis presses the question of whether government by numbers was ever consolidated as a matter of equal protection doctrine, at least in the workplace. An affirmative answer would resolve the tension I posited, but providing that answer requires knowing more about how consolidation occurs and when, under Ackerman’s theory, the window for it closes and that for judicial betrayal opens.

Alternatively, perhaps the tension I identify is illusory because there are simply two constitutional tracks, one embedded in landmark statutes and the other in formal constitutional text. On this account, the civil rights revolution only ever reversed state action under Title VII, leaving intact the traditional doctrine as a matter of enforcing the Constitution’s equal protection provisions. This is the sturdiest solution, as it could also allow government by numbers under Title VII but not equal protection, making easy sense of Davis. It seems to fall far short of Ackerman’s vision for extra-Article V constitutional change, however. He inveighs against the legal profession for ignoring landmark statutes when interpreting Article V-compliant text and urges that they should inform each other.187 Ackerman promises to address in a future volume how to synthesize the civil rights revolution across time, then to now.188 But the fate of government by numbers in the employment discrimination context also calls for elaborating how synthesis should occur synchronically, across the statutory and textual domains of Ackerman’s Constitution. Only by clarifying this dynamic can we truly determine whether the tension I have posited is illusory or real, and whether it is attributable to the consolidation, synthesis, or betrayal of the civil rights revolution.