On June 29, 2023, the Supreme Court ruled in Students for Fair Admissions, Inc. v. President and Fellows of Harvard College, and in the companion case by that same organization against the University of North Carolina (UNC), that the affirmative action woven into the two institutions’ admissions policies violated disfavored applicants’ constitutional rights. Chief Justice John Roberts’s majority opinion, joined by five other Justices, held that “equal protection cannot mean one thing when applied to one individual and something else when applied to a person of another color.” Harvard and UNC’s admissions practices, which are essentially the same as those of all other selective colleges and universities in America, “lack sufficiently focused and measurable objectives warranting the use of race,” Roberts concluded, “unavoidably employ race in a negative manner, involve racial stereotyping, and lack meaningful end points.”

Nearly 28 years before this decision, on July 19, 1995, President Bill Clinton gave a speech at the National Archives building. Standing before display cases that contained the original copies of the U.S. Constitution and Declaration of Independence, Clinton said, “[L]et me be clear: Affirmative action has been good for America.” In response, the audience of Democratic Party luminaries—including cabinet officers, members of Congress, and civil rights leaders—gave the president a loud ovation that lasted for half a minute. The applause was quieter and briefer when Clinton delivered what became the speech’s most famous line: “We should reaffirm the principle of affirmative action and fix the practices. We should have a simple slogan: mend it, but don’t end it.”

For and Against

As journalist Nicholas Lemann noted in The Big Test: The Secret History of the American Meritocracy (1999), which examines how affirmative action and meritocracy’s stories intertwine, the favorable response to Clinton’s speech obscured the fact that it shed no light on what mending affirmative action actually entailed. Strategic ambiguity was decidedly on-brand for the 42nd president. Clinton, now 77, has been an ex-president for almost a quarter century. As a result, few Americans born after 1980 can appreciate the beguiling power of his forensic skills. “The President’s essential character flaw isn’t dishonesty so much as a-honesty,” Michael Kelly wrote in a long 1994 profile for The New York Times Magazine. “Clinton means what he says when he says it, but tomorrow he will mean what he says when he says the opposite.”

Over a long career, the virtuoso feats of a-honesty saw Clinton fashion phrases that appeared to affirm and deny a proposition at the same time, or encouraged two people who disagreed about some issue to believe that he agreed with each of them. Sometimes these constructions were meant to defuse controversies involving Clinton’s personal conduct, such as the claim in 1992 that he had in the past experimented with marijuana but “didn’t inhale.” On other occasions, the purpose was to weigh in on a controversial policy question. Clinton’s position throughout his presidential campaigns and two terms in office, for example, was that abortion should be “safe, legal, and rare.” As The Atlantic’s Caitlin Flanagan pointed out in 2019, he never said how rare abortions should be, or committed himself to any particular government policies designed to reduce their number. Rather, Clinton “located language that made it possible to be completely for legal abortion and against legal abortion.”

At minimum, “safe, legal, and rare” was a Clintonian way to signal that while he was pro-choice, he also respected the pro-life movement’s sincerity in deploring the millions of abortions performed since 1973’s Roe v. Wade decision. Even though this concession was entirely rhetorical, feminists disliked it for validating the idea that abortion was regrettable, which in turn lent a measure of legitimacy to efforts to reduce the number of abortions, including through restrictive laws. Over the past decade, Flanagan points out, “rare” has been “vigorously expunged” from Democratic platforms, Planned Parenthood literature, and the pro-choice vocabulary. In 1992, however, Republicans had held the White House for twelve straight years, during which they had appointed five Supreme Court Justices. On the theory that a fourth consecutive GOP presidential term could mean additional Court nominations that put Roe in even greater jeopardy, defenders of legalized abortion did not rebuke Clinton publicly. Rather, they shared his hope that “safe, legal, and rare” would help him secure an electoral victory that, in turn, fortified and perpetuated the abortion policy regime the Court had created in 1973.

“Mend it, don’t end it” was born of similar political peril. The Democrats had been thrashed in the 1994 midterm elections, becoming the minority party in both chambers of Congress for the first time since 1954. In 1995, activists were gathering signatures for the California Civil Rights Initiative, which became ballot Proposition 209. It proposed to amend the state constitution to prohibit discrimination and preferences based on race, sex, color, ethnicity, or national origin in public contracting, education, and employment. Given that 1994 was already viewed as a repudiation of the Democratic president, journalists had begun writing that the prospect of the 1996 election turning into a referendum on affirmative action would, in Lemann’s words, “portend electoral doom for Clinton.”

Standards of Fairness

In response, “mend it, don’t end it” was meticulously crafted to be as noncommittal as “safe, legal, and rare.” Since mending implies that affirmative action had become flawed or broken in some way, Clinton announced that his administration would enforce “standards of fairness” in superintending affirmative action programs. But upon examination, the standards turned out to rely on distinctions that never identified differences—or to prohibit straw men. Clinton insisted, for example, that he opposed “rigid quotas” but favored the kind of programs that pursued a “flexible goal.” But, of course, if a goal for hiring or matriculating members of an underrepresented minority becomes too flexible—with deadlines that can be postponed over and over, or benchmarks subject to endless renegotiation—it turns into a hope, a wish. If a goal is to effect real change, to “finally address the systemic exclusion of individuals of talent on the basis of their gender or race,” as Clinton said in his speech, then the courts or administrative agencies enforcing it must have some basis for declaring that an organization has too few female employees or a school does not have enough black students.

Not enough, in this circumstance, necessarily means that after every relevant consideration has been accounted for, somebody in authority ultimately divulges what is enough, what number of underrepresented groups in the workforce or student body suffices to remove an employer or school from legal jeopardy. And if having less than X% of a demographic group means that you’re in trouble but having X% or more means that you’re not in trouble, then there’s no reason to insist that it makes some kind of important difference whether we describe X% as a quota or a goal or a target—or a Dodge Durango. The mending of affirmative action that required vigilance against the use of quotas really meant vigilance against outbreaks of candor.

Similarly, Clinton said that he was opposed to any affirmative action program that involved “preference of the unqualified over the qualified of any race or gender.” But affirmative action programs were rarely so blatant, which is why affirmative action opponents did not waste their time contending that innumerate applicants were being admitted to MIT, or tone-deaf ones to Julliard. Their point, rather, was that Congress had passed landmark legislation guaranteeing that no person shall be subjected to discrimination on the basis of race, religion, sex, color, or national origin. The legislators could have employed different terminology. They could, for example, have prohibited discrimination against individuals belonging to a group that, historically, had been excluded from full participation in American public life. Such laws would have lent themselves to the early affirmative action policies that began in the 1960s.

But instead, elected politicians accountable to constituents enacted laws that said no person of any sort shall be subjected to discrimination, which meant that all persons had a right not to be discriminated against on the basis of race, sex, etc. Affirmative action’s opponents have not been zealots for meritocracy. They don’t argue that every college applicant with an SAT score of 1425 deserves to be admitted ahead of any applicant with a 1424. They do insist that when the easy decisions between the qualified and the unqualified are completed, and employers or schools move on to the hard ones about the more and less qualified, that all applicants have a right to expect that their qualifications will be evaluated without reference to the demographic factors excluded by law from consideration.

Instead, the evidence presented by Students for Fair Admissions in their cases against Harvard and UNC showed that demographic categories had a massive role in admissions. Rather than use these considerations as tiebreakers between applicants whose qualifications were closely similar, the two schools operated what were, in effect, separate admissions systems for each demographic group. At Harvard, black applicants who ranked just below average on a scale that combined high school grades with standardized test scores (technically, in the fifth decile, which encompasses scores from the 41st through the 50th percentiles), were twice as likely to be admitted as Hispanic applicants in that decile, and 12 times as likely as Asian-American applicants. For out-of-state students in the fifth decile applying to UNC, blacks were 14 times more likely to be admitted than whites and 33 times more likely than Asian Americans.

“Mending” Without Changing

It turned out, then, that the mending Clinton promised was limited to problems that could be dispelled with hand-waving about the semantic difference between goals and quotas, or by prohibiting only the most flagrant violations against the principle of equal opportunity. Apart from that, the purpose of “mend it, don’t end it” was to mend as little as possible while creating political breathing room for diversity enforcers to continue discriminating against overrepresented groups to whatever extent they felt was necessary, for as long as they deemed it necessary.

In 2016, more than two decades after Clinton’s speech, economist Mark Perry and an informal group of campus watchdogs began filing complaints with the U.S. Department of Education Office of Civil Rights. Their targets were universities that explicitly exclude students from certain lounges in the student union, participation in programs, or other educational activities and benefits on the basis of race or sex. One of Perry’s collaborators notes that the civil rights laws do not have a “good-intentions” carve-out. Though the group has filed nearly 1,000 of these complaints, and seen hundreds resolved to its satisfaction, the fact that higher education is replete with such discrimination, hiding in plain sight, is testament to how little affirmative action has been mended since 1995.

Though “mend it, don’t end it” was ambitious and artful, it was not unique in mischaracterizing affirmative action’s functions and misrepresenting any intentions to revise it. National Review’s Jim Geraghty recently pointed out that Barack Obama sent similar signals, especially as a presidential candidate in 2008. Obama made it clear, for example, that it would be hard to justify racial preferences that boosted the college admissions prospects for his own daughters, both of whose parents graduated from Harvard Law School, over those of, in Obama’s words, “white kids who have been disadvantaged and have grown up in poverty and shown themselves to have what it takes to succeed.”

Such remarks, offered informally rather than in a major speech like the one Clinton delivered, pointed to recasting affirmative action to make it class- rather than race-based. It is a position most closely associated with Richard Kahlenberg, author of The Remedy: Class, Race, and Affirmative Action (1996). The idea is that it is better to help people who need it by assessing need directly than to use race, sex, or ethnicity as a proxy for need. The resulting policy will be more efficient, conferring help on people to whom it will make a big difference, rather than on those to whom it won’t, such as Sasha and Malia Obama. The politics will be better, too. “If you want working-class white people to vote their race,” according to The Remedy, “there’s probably no better way to do it than to give explicitly racial preferences in deciding who gets ahead in life.”

Whatever Obama’s true feelings on the subject, he emulated Clinton by spending eight years in the White House without making even slight modifications to the way affirmative action programs function in America. In January 2023, Kahlenberg wrote in the Washington Monthly that 71% of the black, Hispanic, and Native American students at Harvard come from families in the top socioeconomic quintile of each of those groups. Ten percent of blacks in America are first- or second-generation immigrants but, according to Kahlenberg, 40% of black students in the Ivy League come from such families, and are disproportionately “likely to be the children of highly educated and wealthy parents.”

Baiting, Then Switching

Bill Clinton’s final standard of fairness for mending affirmative action was that “as soon as a program has succeeded, it must be retired.” The word choice conveys urgency: breathless aides bursting into the Oval Office to inform the president that, somewhere, an affirmative action program had finally achieved its flexible goal, thus requiring the government to deploy every resource at its disposal to decommission that endeavor, preferably by the close of business that very day.

Of course, no affirmative action programs were retired at any point during Clinton’s two terms in office. Connecting the key points of his speech, Clinton’s position amounted to saying that affirmative action had been good for America, so must be ended as quickly as possible. From the start, affirmative action’s defenders had made a point of portraying it as a temporary departure from the standard of treating and judging every American as an individual, without conferring benefits or imposing penalties because someone happened to be of a certain race, sex, or nationality. In his dissent to the 1978 Regents of the University of California v. Bakke decision, Justice Harry Blackmun voted to uphold the University of California, Davis medical school’s policy of excluding white applicants from 16% of its admissions slots, while voicing the “earnest hope” that the time when affirmative action becomes unnecessary might be reached “within a decade at the most.” Twenty-five years later, in her majority opinion in Grutter v. Bollinger (2003), Justice Sandra Day O’Connor expressed the same aspiration but on a different schedule: “We expect that 25 years from now, the use of racial preferences will no longer be necessary” for selective law schools to enroll an entering class with the desired degree of racial heterogeneity. In 2008 Barack Obama also allowed that “affirmative action is not going to be the long-term solution to the problems of race in America.” Little wonder that Chief Justice Roberts concluded in the Harvard and UNC opinion that affirmative action programs “lack meaningful end points.”

It would have been refreshing if more arguments on behalf of affirmative action had forthrightly explained that it was not a brief detour in the journey to equal rights and standing under the law, but a long road that America had no choice but to travel. In Bakke, the four liberal justices who endorsed the U.C. Davis program argued that it was “consistent with the goal of putting minority applicants in the position they would have been in if not for the evil of racial discrimination.” The logic underwhelms: counterfactual speculation cannot, in fact, ascertain sociological details about the alternate-universe United States that would exist if chattel slavery and Jim Crow had never occurred. This audacious epistemology in the service of audacious social engineering was, however, candid about affirmative action’s sweeping implications. For all that, the Justices’ statement was a single line that appeared in a footnote, a point that was never elaborated by other prominent political leaders, and one that played no role in the subsequent debates over affirmative action.

Affirmative action’s defenders chose to dissemble about how it functioned and how long it would last because they faced two related problems. First, the landmark civil rights laws of the 1960s had been presented to the country with fervent pledges that the laws would not promote—would, in fact, categorically prohibit—the use of flexible goals to counteract the disparate impact of neutral standards. The Leadership Conference on Civil Rights, for example, released a statement insisting that the 1964 Civil Rights Act being debated in Congress would not require employers “to maintain any kind of racial or religious balance.” “Indeed, preferential treatment of Negroes or any minority group would violate this section. Its whole point is that all workers should be treated equally.”

Following complaints that the bill, as written, would outlaw employment tests if there were racial disparities in applicants’ results, the two senators managing its progress toward passage, Democrat Joseph Clark of Pennsylvania and Republican Clifford Case of New Jersey, distributed a memorandum insisting that the possibility was a red herring:

There is no requirement in title VII that employers abandon bona fide qualification tests where, because of differences in background and education, members of some groups are able to perform better on these tests than members of other groups. An employer may set his qualifications as high as he likes, he may test to determine which applicants have these qualifications, and he may hire, assign, and promote on the basis of test performance.

As soon as the laws were enacted, of course, officials administering them began to construe their provisions as though the guarantees that individual rights would never be sacrificed to group rights meant the opposite of what the legislation, and the legislators, had said. It’s fair to ask, though impossible to know, whether the politicians who did the baiting in order to get the legislation passed anticipated or even welcomed the switching done by the bureaucrats and jurists after the bills were signed. The fact that the first group, the people who gave the assurances, never denounced the second group, the ones who nullified those assurances, is more than slightly suspicious.

The Wrong Fight

What is clear is that the guarantees about civil rights laws’ strict limits, like Bill Clinton’s circumlocutions about mending practices that were never to be ended, were based on a second problem: affirmative action has never been popular. The rhetoric in the 1960s promising adherence to the principle of equal individual rights was delivered in anticipation of this problem. The subsequent rhetoric minimizing and misrepresenting how far affirmative action had departed from this principle was delivered in response to it.

It is surprising, in retrospect, that it took affirmative action’s opponents 30 years to find a way to let voters weigh in on the question. When they did, Californians approved Proposition 209 and rejected affirmative action. (Clinton hedged his bets throughout. Lemann reports that he waited until the final week of the 1996 presidential campaign, after it had become clear that he would win an easy victory against Bob Dole, to say publicly that he opposed 209.) By 2020, when California voters by even larger margins rejected a ballot proposition to repeal 209, eight other states had prohibited affirmative action, most by referendum, some by legislation.

Public opinion surveys have also been clear and consistent: large majorities of Americans oppose the use of preferential treatment or consideration as a measure to help underrepresented minority groups. In 15 surveys taken by the Pew Research Organization from 1987 to 2012, the proportion of respondents who disagreed with the statement, “We should make every possible effort to improve the position of blacks and other minorities even if it means giving them preferential treatment,” ranged from 62% to 72%. The numbers who agreed were between 24% and 34%.

An Economist/YouGov poll taken the week after the Supreme Court’s 2022-23 term found that 59% of respondents approved of the Court’s decision against the constitutionality of affirmative action in college admissions: 46% strongly approved and 13% somewhat approved. Twenty-seven percent disapproved, 18% strongly and 9% somewhat. Whites approved 65% to 23%. So did Hispanics, 45% to 30%. And so did blacks: 31% strongly and 13% somewhat, compared to 26% who strongly disapproved and 10% who somewhat disapproved. (This result likely bears some relation to the finding that only 11% of blacks said that affirmative action had affected them positively.)

The poll also argues that Americans prefer meritocracy, narrowly defined, over identity politics, at least in college admissions. Given a list of factors colleges should be allowed to consider as part of the admissions process, 85% of all respondents favored high school grades and 77% endorsed scores on standardized tests. Meanwhile, 73% (of both men and women) opposed consideration of gender. Sixty-nine percent opposed consideration of race or ethnicity: 74% of whites, 60% of Hispanics, and 54% of blacks.

One protagonist in Nicholas Lemann’s The Big Test is Molly Munger, a wealthy Californian (the daughter of Warren Buffett’s business partner) who was a Democratic Party donor and activist. In 1996 she became a central figure in the campaign to save affirmative action by defeating Prop. 209. After pondering the results from Election Day, which saw Clinton carry California but 209 win an even larger majority, Munger, in Lemann’s summary, “lost faith in…affirmative action.”

If people who cared about social justice, especially racial justice, built their cause around affirmative action, she feared they were destined to lose and lose and lose in perpetuity. It was the wrong fight to be in. The right fight was the fight to make sure that everybody got a good education and a chance to live a life of decency and honor, and that the country held together as a community.

Democrats never could bring themselves to mend affirmative action—they rarely attempted even cosmetic changes. If the Students for Fair Admissions decisions mean that the Roberts Court is prepared to end affirmative action, it may wind up extricating Democrats from a dilemma they could not solve on their own. We can hope that insistence on the principle that no persons are more equal than others redirects the party, and the country, to an honest, productive debate over opportunity, honor, and national cohesion.