From Bakke to Grutter, the Long Legal Life of Affirmative Action

Allan Bakke was a former Marine officer and NASA engineer in his thirties when he applied to medical school at the University of California, Davis, in the early 1970s. He was rejected twice. The Davis medical school, which opened in 1968, set aside 16 of the 100 seats in each entering class for a special admissions program for disadvantaged minority applicants, who were evaluated separately from everyone else. Bakke, who was white, sued, arguing that he had been shut out of seats he could never compete for because of his race. When his case reached the Supreme Court in 1978, the justices split so badly that no majority agreed on a rationale. The opinion that ended up governing American admissions for a generation was written by one justice, Lewis Powell, speaking mostly for himself. That fact is the key to everything that followed.

In short: In Regents of the University of California v. Bakke (1978), Justice Powell's lone controlling opinion struck down racial quotas but allowed race as one factor in admissions to serve educational diversity. The Court reaffirmed that idea in Grutter v. Bollinger (2003) and then abandoned it in the 2023 SFFA decisions, ending a legal theory that had always rested on a narrow and contested foundation.

Where affirmative action started

The phrase "affirmative action" entered federal policy in 1961, when President John F. Kennedy issued Executive Order 10925 requiring government contractors to take affirmative action to ensure that applicants and employees were treated without regard to race, creed, color or national origin. In 1965 President Lyndon Johnson replaced it with Executive Order 11246, which became the backbone of federal contractor rules for six decades until it was revoked in January 2025.

Johnson also made the moral case in a commencement address at Howard University in 1965, arguing that it was not enough to open the gates of opportunity; people who had been held back for generations needed the ability to walk through them. Colleges, many of which had been all or nearly all white, began actively recruiting Black students in the late 1960s. Some set targets or reserved seats. The Davis program was one of these, and it was candid in its structure: a separate committee, a separate track, a fixed number.

An earlier challenge, DeFunis v. Odegaard, reached the Court in 1974 from the University of Washington law school, but the justices dismissed it as moot because the plaintiff, who had been admitted under a lower-court order, was about to graduate. Bakke was the first real test.

The Bakke split, counted carefully

The 1978 decision is easiest to understand as two separate 5 to 4 votes.

On the first question, whether the Davis program was illegal and Bakke had to be admitted, five justices said yes. Four of them (led by Justice John Paul Stevens) said the program violated Title VI of the Civil Rights Act, which bars race discrimination by programs receiving federal funds, and they did not reach the broader constitutional question. Powell joined them in the result but on constitutional grounds: a fixed racial quota violated the Equal Protection Clause.

On the second question, whether a university could ever consider race at all, five justices said yes. Four of them (led by Justice William Brennan) would have upheld the Davis program entirely, arguing that remedying the effects of past societal discrimination was a valid reason for race-conscious measures. Powell joined them in the result but rejected their rationale.

So Powell was the swing on both halves, and his reasoning, which no other justice fully joined, became the law that lower courts and universities followed. Justice Thurgood Marshall wrote a separate opinion recounting the history of slavery and segregation and arguing that the country could not now hide behind formal neutrality. Justice Harry Blackmun wrote that to get beyond racism, we must first take account of race, a line that became one of the best known of the case.

Why Powell chose diversity

Powell rejected most of the reasons Davis offered. Remedying general societal discrimination was too broad and unbounded, he thought; a university was not a court and could not make findings about who had been wronged. Increasing the number of minority doctors, or improving health care in underserved communities, might be worthy, but Davis had not shown its program was needed to achieve them.

He accepted one reason: the educational benefits of a diverse student body. A university, he wrote, has a First Amendment interest in academic freedom that includes choosing whom to teach, and a mix of students with different backgrounds and viewpoints enriches the training of everyone. Race could be one "plus" factor among many, considered alongside geography, life experience, talents and so on. As his model he pointed to the admissions program at Harvard College, which considered race as one element in a flexible review.

That choice had consequences that took decades to play out. The diversity rationale was forward-looking and benefited everyone, which made it easier to defend politically. But it cut the policy loose from its original purpose, repairing harm to Black Americans. It invited the question of why race and not other kinds of diversity. And because it relied on the idea that every applicant was being weighed individually, it made universities defend their programs as holistic and flexible, while the numbers they produced often looked very stable year to year. That gap between theory and practice was exactly where later challengers aimed.

There is a bit of irony buried in the record, too. The Harvard plan that Powell held up as a model in 1978 was the program that, 45 years later, the Court struck down.

Hopwood, Proposition 209 and the states

In the 1980s and early 1990s, the Court moved against race-conscious measures in other areas. In City of Richmond v. J.A. Croson (1989) it struck down a city set-aside for minority-owned contractors, and in Adarand Constructors v. Peña (1995) it held that federal racial classifications must meet strict scrutiny, the highest level of review.

In 1996 the federal appeals court covering Texas ruled in Hopwood v. Texas that the University of Texas law school could not use race in admissions at all, declaring that diversity was not a compelling interest. That same year, California voters passed Proposition 209, banning racial preferences in public education, employment and contracting. Washington and, later, Michigan and several other states followed with bans of their own.

So by the turn of the century, affirmative action was alive in federal constitutional law but dead in some of the largest public university systems in the country. Its survival nationally depended on the Supreme Court reaffirming Powell's lonely opinion.

Grutter, Gratz and Justice O'Connor's 25 years

It did, in 2003, in a pair of cases from the University of Michigan.

In Grutter v. Bollinger, decided 5 to 4 with Justice Sandra Day O'Connor writing for the majority, the Court upheld the law school's admissions program. This time a majority, not just one justice, adopted the diversity rationale and held that it was a compelling interest. The law school's program, which considered race flexibly and sought a "critical mass" of underrepresented students, was narrowly tailored, the Court said. O'Connor added a famous sentence of expectation: the Court expected that 25 years from then, the use of racial preferences would no longer be necessary to further the interest approved that day.

On the same day, in Gratz v. Bollinger, the Court struck down Michigan's undergraduate program 6 to 3 because it automatically awarded a fixed number of points, 20 on a 150-point scale, to applicants from underrepresented minority groups. A mechanical bonus, the Court said, was too much like a quota.

The two rulings, read together, gave universities a clear instruction: individualized review was allowed, formulas were not. It also pushed them toward opacity. The more a college could say that race was one subtle element in a holistic judgment, the safer it was legally, even though that made it harder for anyone, including applicants, to know how much race actually mattered.

Many supporters of affirmative action were relieved by Grutter and uneasy at the same time. The dissenters, including Chief Justice William Rehnquist and Justices Antonin Scalia, Clarence Thomas and Anthony Kennedy, made clear they thought the program was a disguised quota. Kennedy's view in particular would matter later.

The generals and the corporations who showed up

One reason Grutter came out the way it did was who filed briefs in support of Michigan. Alongside the expected civil rights groups and universities, a group of retired senior military officers and service academy leaders told the Court that a racially diverse officer corps was essential to national security. They pointed to the Vietnam era, when a mostly white officer corps commanded enlisted ranks with large numbers of Black troops, and described the breakdown in trust and discipline that followed. The academies and ROTC programs, they argued, needed to consider race to produce officers who looked like the forces they led.

Dozens of major corporations filed briefs too, arguing that their workforces and customers were diverse and that they needed graduates who had learned alongside people from different backgrounds.

O'Connor's opinion relied on both. The arguments moved the case out of the frame of a contest between minority applicants and white applicants and into a frame of national institutions saying what they needed. It was a smart coalition, and it worked for twenty years.

It also showed the limits of the diversity frame. When the institutions changed their minds, or when the Court changed its membership, the coalition's argument was only as strong as the justices' willingness to defer to it. In 2023 the majority set the military academies aside in a footnote, leaving that question for another day, and did not find the corporate arguments sufficient for civilian universities. The lesson about coalitions is double-edged: broad support can win a case, but it cannot hold a precedent in place by itself.

Fisher, Schuette and the slow erosion

Over the next decade the composition of the Court changed. O'Connor retired in 2006 and was replaced by Samuel Alito.

In Fisher v. University of Texas (2013), the Court sent a challenge back to the lower courts with instructions to apply strict scrutiny more rigorously. In Schuette v. BAMN (2014), it upheld Michigan's voter-approved ban on affirmative action, holding that voters could choose to end such programs. Then in the second Fisher case in 2016, Kennedy, who had never before voted to uphold a race-conscious admissions program, wrote a 4 to 3 opinion upholding the University of Texas's plan, while warning that universities had an ongoing duty to show that race-neutral alternatives would not work.

That was the last time affirmative action won at the Court. With Kennedy's retirement in 2018 and the confirmation of three new justices between 2017 and 2020, the majority that had sustained Grutter was gone. Students for Fair Admissions, the group led by the longtime affirmative action opponent Edward Blum, had already sued Harvard and UNC in 2014. When the cases reached the Court, a six-justice majority held in 2023 that both programs were unlawful. The 25-year expectation O'Connor expressed in 2003 had been overtaken at about the 20-year mark.

Why the foundation was always fragile

Here is the argument this history supports. Progressives spent four decades defending race-conscious admissions on the ground that Powell gave them, diversity, partly because it was the only ground the Court would accept. That made legal sense, but it carried two costs.

It made the policy look like something elite institutions did for their own educational benefit, rather than a remedy for specific, documented harms done by public institutions, including discrimination in housing, schools and lending that government itself had organized. And because the diversity rationale applied to everyone, it was hard to explain to the public why it required race in particular, or when it would end.

The opponents' strongest case, stated fairly, was consistent from Bakke onward: the Constitution and the Civil Rights Act protect individuals, not groups, and any racial classification by government invites resentment and stereotype, whatever its motive. Many Americans of all races found that persuasive, and it is the view that eventually commanded a majority of the Court.

The answer is not to pretend that view has no merit. It is to point out that the country has never actually been race-neutral in the decisions that shaped who could build wealth, buy a home in a good school district, or attend a well-funded high school. The policy debate now has to move to those upstream decisions, where many remedies, like school funding reform, need-based aid and fair housing enforcement, do not require racial classification at all. Our racial justice page tracks that work.

What you can do

Bakke turned on who sat on the Court and how they reasoned. That remains true.

1. Watch who gets confirmed. Supreme Court opinions this term, with the Court's own summaries, plus Senate votes on judges, are all on Court Watch. Lower-court judges decide most of these questions first, and the Senate votes on every one.
2. Call your senators about judicial nominees. Find your representatives by address to get their phone numbers, contact forms and their latest votes. Senators hear from constituents on judges far less often than on spending, which gives each call more weight.
3. Follow the courts beyond this one issue. Our courts page gathers reporting on the federal judiciary, from the emergency docket to the circuit courts where most cases end.
4. Support the upstream fixes. Groups working on school funding equity, college access and fair housing are listed by state in the organizations directory. Their work reaches far more students than any admissions office.