In McCleskey v. Kemp, decided April 22, 1987, the Supreme Court ruled 5–4 that statistical evidence of racial bias in capital sentencing, no matter how strong, cannot by itself invalidate a death sentence. A defendant has to prove that the prosecutor, judge, or jurors in his own case acted with discriminatory intent. Warren McCleskey, a Black man sentenced to death in Georgia for killing a white police officer, brought the challenge with what remains one of the most detailed studies of racial disparity in death sentencing ever assembled. The Court accepted the numbers and ruled against him anyway.
The Case Behind the Ruling
In 1978, McCleskey and three accomplices robbed a furniture store in Atlanta. A police officer responding to a silent alarm was shot and killed during the robbery. A Georgia jury convicted McCleskey of armed robbery and murder and recommended death on the murder count, and the trial court imposed the sentence.1Legal Information Institute. McCleskey v Kemp
After the Georgia Supreme Court affirmed, McCleskey sought federal habeas relief. His central evidence was a statistical study of Georgia capital cases. The federal district court and the en banc Eleventh Circuit both rejected his claim; the Supreme Court took the case to decide whether that evidence could carry a constitutional challenge.2Justia. McCleskey v Kemp, 481 US 279 (1987)
The Baldus Study
Professor David Baldus and his colleagues reviewed more than 2,000 Georgia murder cases from the 1970s. They tracked the race of the defendant, the race of the victim, and 230 nonracial variables that might legitimately affect a sentence: the severity of the crime, the defendant’s criminal history, the strength of the evidence, and so on.1Legal Information Institute. McCleskey v Kemp
The numbers were striking. The death penalty was imposed in 22 percent of cases involving Black defendants and white victims, compared to 3 percent of cases involving white defendants and Black victims. Prosecutors sought death 70 percent of the time when a Black defendant was charged with killing a white victim, and 19 percent of the time when a white defendant was charged with killing a Black victim. Once every nonracial variable was controlled for, defendants whose victims were white were 4.3 times more likely to receive a death sentence than defendants whose victims were Black.1Legal Information Institute. McCleskey v Kemp
The disparities concentrated in the middle range of cases. The most brutal murders produced death sentences across the board. The least serious cases did not. In the gray zone, where the facts alone left the outcome genuinely open, race exerted the strongest pull.
The Two Constitutional Arguments
McCleskey’s lawyers pressed two theories. The Eighth Amendment claim built on Furman v. Georgia (1972) and Gregg v. Georgia (1976): Furman had struck down capital sentencing schemes because they created a substantial risk of arbitrary punishment, and Gregg had upheld Georgia’s rewritten scheme because it channeled discretion. If the race of the victim could predict a death sentence better than many legitimate factors, McCleskey argued, the arbitrariness Furman targeted was still there, just in a different form.2Justia. McCleskey v Kemp, 481 US 279 (1987)
The Fourteenth Amendment claim was that the state was administering capital punishment in a racially discriminatory way. Rather than pointing to a specific biased actor, McCleskey argued that operating a system known to produce racially skewed outcomes should itself count as intentional discrimination, or at least shift the burden to the state to explain why race was not driving the results.1Legal Information Institute. McCleskey v Kemp
What the Court Held
Justice Lewis Powell wrote for a five-justice majority joined by Chief Justice Rehnquist and Justices White, O’Connor, and Scalia. The Court rejected both claims.
On equal protection, the majority held that McCleskey had to prove the decisionmakers in his own case acted with discriminatory purpose. He offered nothing specific to his trial, and the Baldus study alone could not support the inference that any particular juror or prosecutor was motivated by race. The Court acknowledged that statistics have long been accepted as proof of discrimination in jury selection and in Title VII employment cases, but drew a line at capital sentencing, calling it “fundamentally different” because each sentence is an individualized decision by a jury weighing many factors about a specific crime and defendant.2Justia. McCleskey v Kemp, 481 US 279 (1987)
On the Eighth Amendment, the Court held that because Georgia’s post-Gregg procedures channeled discretion toward the individual crime and defendant, McCleskey’s sentence could be presumed non-arbitrary. Statistical disparities across thousands of cases did not make the system “completely arbitrary and lacking in reason.”2Justia. McCleskey v Kemp, 481 US 279 (1987)
An institutional worry ran through the opinion. Justice Powell wrote that if the Court accepted McCleskey’s theory, it could “soon be faced with similar claims as to other types of penalty,” potentially throwing “into serious question the principles that underlie our entire criminal justice system.” The opinion did not deny that racial disparities might exist. It treated the consequences of remedying them as too destabilizing to accept.1Legal Information Institute. McCleskey v Kemp
The Dissents
Justice Brennan, joined by Justices Marshall, Blackmun, and Stevens, called the majority’s institutional worry a “fear of too much justice.” Since Furman, the Eighth Amendment analysis had focused on the risk of arbitrary sentencing, not on proven arbitrariness in any single case. The Baldus study demonstrated that risk with what Brennan called “unprecedented refinement and strength.”2Justia. McCleskey v Kemp, 481 US 279 (1987)
Justice Blackmun, joined by Marshall and Stevens, attacked the majority’s use of discretion as a defense. Discretion exists so jurors can treat each defendant as an individual, not so they can sort people by race. An enhanced willingness to impose death on Black defendants, or a diminished willingness when the victim is Black, “reflects a devaluation of the lives of black persons.” Blackmun also pointed out that Georgia gave prosecutors and juries essentially unstructured discretion, with no guidelines governing the decision to seek death.2Justia. McCleskey v Kemp, 481 US 279 (1987)
Powell’s Later Regret
Warren McCleskey was executed in Georgia’s electric chair on September 25, 1991. He was 44.
Years later, Justice Powell’s biographer John Jeffries asked whether he regretted any of his votes. Powell named McCleskey. Pressed on whether he had come to favor statistical claims of systemic bias, Powell said no. He had concluded capital punishment itself should be abolished. The justice who wrote the majority opinion no longer believed the system he defended could be made fair.
Why the Ruling Still Matters
McCleskey effectively closed the federal courthouse to claims of racial bias in sentencing based on statistical evidence. A defendant can still bring an equal protection challenge, but only by proving that a specific prosecutor, judge, or juror in the case acted with discriminatory intent. That is a difficult showing in a system where bias more often operates through unconscious assumptions and accumulated discretionary choices than through overt statements. Legal scholars have compared the decision to Dred Scott v. Sandford for its consequences for racial justice, and it remains among the most criticized rulings in modern constitutional law.
The decision also carved out capital sentencing from areas where statistical proof is routine. Courts still accept patterns of racial disparity as evidence in employment discrimination and jury-selection cases. In the setting with the highest stakes, the Court held statistics alone were not enough.2Justia. McCleskey v Kemp, 481 US 279 (1987)
State Racial Justice Acts
Because McCleskey pushed remedies out of the courts and toward legislatures, a few states have passed statutes creating what federal law does not: a way to challenge racially biased sentencing using statistical evidence.
Kentucky went first in 1998. Its Racial Justice Act bars a death sentence where clear and convincing evidence shows the prosecutor’s decision to seek death was based on race. The law applies only before trial, cannot be raised retroactively, and does not reach jury selection or the sentencing phase. No Kentucky capital defendant has succeeded under it.
North Carolina enacted a broader statute in 2009, allowing death-sentenced prisoners to challenge their sentences by showing race played a role in sentencing or jury selection. The legislature later repealed the act, and the repeal itself became the subject of litigation before the North Carolina Supreme Court.
California passed the most expansive version in 2020. Codified at Penal Code Section 745, the California Racial Justice Act lets defendants use statistical evidence of racial disparities in charging, convictions, and sentencing within their county. It does not require proof that the bias was intentional, reaching both explicit and implicit prejudice. As of January 1, 2026, the law applies retroactively to anyone with a felony conviction or a juvenile case that resulted in commitment.3Office of the State Public Defender. Racial Justice Act Retroactivity
These statutes remain the exception. Most states have no Racial Justice Act, and the McCleskey standard still governs federal constitutional claims of racial bias in sentencing across the country.