Justice Kennedy’s Criminal Justice Legacy

David Patton

When Justice Sandra Day O’Connor’s retired in 2005, Justice Anthony Kennedy became the undisputed swing vote on the Supreme Court across a range of topics. But even with that title as a baseline, Kennedy’s vote was particularly critical on matters of criminal justice. Consider the numbers: In the 10 years after O’Connor’s retirement, in 5–4 decisions in which the defendant prevailed, Kennedy provided the fifth vote along with the four liberal members of the court an extraordinary 27 times (triple the number of the next most frequent conservative justice to cast a pivotal pro-defendant vote, Justice Scalia). (Michael A. McCall & Madhavi M. McCall, Quantifying the Contours of Power: Chief Justice Roberts & Justice Kennedy in Criminal Justice Cases, 37 Pace L. Rev. 115, 151–55 (2016).) That’s not to say that Kennedy always swung to the defendant; indeed, in roughly equal measure, he provided the decisive vote for the government. One way to think about Kennedy’s criminal justice jurisprudence is to imagine him at a dinner party in Paris wanting to make a good impression with his companions. What about the death penalty, Justice Kennedy? Ah, barbaric so much of the time. Don’t you think your legal system should pay more attention to international norms? Mais oui! Can you explain this American obsession with juries? Sometimes it’s beyond me. And isn’t your system awfully adversarial? Truth be told, it can be a bit much for my liking.

Below is an analysis of Kennedy’s impact on some of the most notable criminal justice issues of his era, and how the Court may respond in his absence. Generalizations about justices are often dangerous, but it seems safe to say that there were two broad divides that were good (though not perfect) predictors for Kennedy’s leanings: capital and noncapital, and juvenile and adult. He tended to favor defendants in the former categories and the government in the latter. He greatly narrowed the class of persons subject to the death penalty and the amount of punishment that could be imposed upon juveniles, but in the noncapital, adult arena, he could be a pro-prosecution vote in some of the most significant cases of his time, particularly on a range of Fourth Amendment issues, and the line of Sixth Amendment cases defining the proper role of the jury at sentencing and the importance of the Confrontation Clause.

As discussed below, Kennedy’s legacy, whether pro-defendant or pro-government, and how durable it will be, varies considerably from topic to topic. The durability, of course, depends on the votes of his replacement, Justice Brett Kavanaugh, as well as the behavior of Chief Justice Roberts, the likely new center of the Court. Whether Roberts changes his voting behavior as a result of his new role in the middle, and how strictly he adheres to stare decisis, will likely determine the lasting impact of Kennedy’s legacy.

The Death Penalty

Beyond any doubt, Kennedy’s most profound impact on the criminal justice system was his repeated fifth vote for defendants in capital cases. Here, more than any other area, Kennedy lived up to the swing vote label. And he did so in the traditional sense of joining with the four liberal members of the Court over the constant, vigorous dissents of the other four conservative members. In the most consequential capital cases of the past 10 years, the dissenting line-up of Roberts, Scalia, Thomas, and Alito became a broken record. And even before then, when O’Connor was still on the Court, Kennedy began to swing left. The cases are now profound parts of the jurisprudence of the death penalty: Atkins, Roper, Kennedy, and Hall, among others.

In 2002 in Atkins v. Virginia, O’Connor and Kennedy joined the liberal wing to hold that the Eighth Amendment prohibits the execution of anyone who was intellectually disabled at the time of the commission of the capital offense. (536 U.S. 304 (2002).) Three years later, this time with O’Connor in dissent, Kennedy authored a 5-4 decision in Roper v. Simmons, holding the same with respect to juveniles: Persons who were under the age of 18 at the time they committed a capital offense may not be executed. (543 U.S. 551 (2005).) Roper is notable in a number of respects. First, with O’Connor still on the Court, it revealed Kennedy as the true swing vote on capital cases. Second, it overturned a decision that was only 15 years old, a decision from 1989 when Kennedy himself was on the Court and voted in the opposite direction. (Stanford v. Kentucky, 492 U.S. 361 (1989).) And lastly, Kennedy’s opinion relies upon sources that draw scorn from the right side of the Court: scientific research into the brain development of juveniles and, most unforgivably, the international consensus against the death penalty for juveniles, citing the United Nations Convention on the Rights of the Child and the United Kingdom’s history of abolition. With typical rhetorical flourish (of the sort that brings ridicule from the right), Kennedy explained:

It does not lessen our fidelity to the Constitution or our pride in its origins to acknowledge that the express affirmation of certain fundamental rights by other nations and peoples simply underscores the centrality of those same rights within our own heritage of freedom.

(Roper, 543 U.S. at 578.) The decision aroused fury from Justice Scalia:

What a mockery today’s opinion makes of Hamilton’s expectation [about the role of judges], announcing the Court’s conclusion that the meaning of our Constitution has changed over the past 15 years—not, mind you, that this Court’s decision 15 years ago was wrong, but that the Constitution has changed. . . . The Court [] proclaims itself sole arbiter of our Nation’s moral standards—and in the course of discharging that awesome responsibility purports to take guidance from the views of foreign courts and legislatures.

(Id. at 608 (Scalia, J., dissenting).) The enraged reaction from Scalia immediately became a rallying cry for Federalist Society conservatives, even prompting Republican legislative proposals to ban Supreme Court citations to foreign authorities. (Jess Bravin, Congress May Fight Court on Global Front, Wall St. J., Mar. 21, 2005.)

Following Roper, the list of 5-4 decisions in favor of defendants in capital cases, in which Kennedy provided the fifth vote and authored the majority opinion against a solid wall of conservative dissent, is breathtaking: Smith v. Texas (550 U.S. 297 (2007)), holding that Texas erred in not permitting the jury to adequately consider mitigating factors, that the issue was properly preserved, and that the error was not harmless; Panetti v. Quarterman (551 U.S. 930 (2007)), holding that a prisoner must have a “rational understanding” of the reason for his execution at the time of his execution to be competent under the Eighth Amendment; Kennedy v. Louisiana (554 U.S. 945 (2008)), holding that the death penalty is impermissible under the Eighth Amendment for the nonhomicide crime of child rape; and Hall v. Florida (572 U.S. 701 (2014)), expanding the holding in Atkins and finding that Florida’s rule disallowing exploration of intellectual disability to anyone with an IQ score above 70 created “an unacceptable risk that persons with intellectual disabilities will be executed.” (Id. at 704.)

And though he was not the author, he was the indispensible fifth vote in several other notable 5-4 decisions: Brewer v. Quarterman (550 U.S. 286 (2007)) and Abdul-Kabir v. Quarterman (550 U.S. 233 (2007)), in which Kennedy joined Justice Stevens’ opinions that Texas erred in not allowing jurors to meaningfully and fully consider mitigating factors; and Brumfield v. Cain (135 S. Ct. 2269 (2015)), authored by Sotomayor, holding that the state court’s determination that an IQ score of 75 meant that the prisoner could not meet the Atkins test for intellectual disability was error and that the prisoner was entitled to have his Atkins claim considered on the merits in federal court.

The most prominent topic on which Kennedy was a disappointment to capital defendants was lethal injection. In Baze v. Rees (553 U.S. 35 (2008)) and Glossip v. Gross (135 S. Ct. 2726 (2015)), Kennedy joined the conservatives in finding that the three-drug protocol had not been shown to be cruel and unusual. Kennedy’s vote was particularly important in Glossip, a 5–4 decision with all of the liberals in dissent.

For the same reason that Kennedy’s impact on the death penalty was extraordinary, it is also fragile. His votes were pivotal, and his replacement in Kavanaugh, a jurist hand-picked by the Federalist Society, surely means that the cases above (lethal injection aside) rest on thin ice. At the very least, the odds of the Court expanding the holdings have gotten longer. Obvious open questions at the time of his retirement included whether Roper’s prohibition on execution might extend past the age of 18 (as brain science suggests it should), whether Atkins’s definition of intellectual disability and the means of demonstrating it should broaden; and whether Kennedy v. Louisiana might ultimately lead to a prohibition on the death penalty for nonhomicide crimes against the state (in addition to the prohibition for crimes against individuals). And, of course, the odds also got much longer for the ultimate prize: complete abolition.


Justice Kennedy had an extraordinary impact on the sentencing of juveniles. In addition to Roper, which was pathbreaking in its acknowledgment and discussion of the fundamental, scientific distinction between juveniles and adults, he wrote the majority opinion in Graham v. Florida, categorically prohibiting sentences of life imprisonment without parole for juveniles convicted of nonhomicide offenses and further requiring that those juveniles receive a “meaningful opportunity for release.” (560 U.S. 48, 50 (2010).) Although the Court’s judgment vacating Graham’s sentence received six votes, Roberts did not join the majority opinion because he only found the particulars of Graham’s sentence to violate the Eighth Amendment and would not have created a categorical ban on such sentences. Thus, Kennedy again provided the critical fifth vote for a new and sweeping Eighth Amendment protection.

He did so once more in Miller v. Alabama. (567 U.S. 460 (2012).) Although he did not author the opinion, he was again the sole conservative to join the liberal wing in holding that the Eighth Amendment prohibited a sentence of mandatory life imprisonment for juveniles convicted of any crime, including homicide. The 5–4 decision, later made retroactive in the Kennedy-authored opinion in Montgomery v. Louisiana (136 S. Ct. 718 (2016)), meant new opportunities for eventual release for over 2,500 juveniles nationwide who were sentenced to mandatory life without parole sentences.

There is a ray of hope for the durability of Miller beyond Kennedy’s tenure. Roberts voted to apply its holding retroactively in Montgomery—despite a vigorous dissent from Scalia, Thomas, and Alito on traditional conservative grounds relating to finality and deference to state court decisions. It may overstate the significance of Roberts’s vote, but perhaps, despite his disagreement with Miller and Graham, he has come to accept them as settled law.

Another notable fifth vote from Kennedy with respect to juveniles was his agreement with the four liberal justices in J.D.B. v. North Carolina, that the age of a child being interrogated by police is relevant to the question of whether the child is in custody for purposes of a Miranda analysis. (564 U.S. 261 (2011).) Kennedy was not typically on the side of defendants in procedural matters relating to police interrogations; indeed he was the critical fifth vote and author of decisions in cases like Berghuis v. Thompkins, 560 U.S. 370 (2010) and Yarborough v. Alvarado, 541 U.S. 652 (2004) which greatly curtailed Miranda rights. But in J.D.B., he was once again moved by the juvenile status of the defendant.

The Sixth Amendment and Juries

Many commentators consider the Court’s jurisprudence over the past 20 years with respect to the proper role of juries and judges in determining sentences nothing short of revolutionary. It is hard to disagree. But if it was indeed a revolution, Kennedy was part of the palace guard.

Beginning with Apprendi v. New Jersey in 2000, the Court fundamentally enhanced the role of juries in determining sentences. (530 U.S. 466 (2000).) In Apprendi, a 5–4 decision authored by Stevens and joined from the right by Scalia and Thomas, with Kennedy in dissent, the Court held that any fact other than the fact of a prior conviction that increases a penalty for a crime beyond the statutory maximum must be submitted and proved to a jury beyond a reasonable doubt.

Within six years, the full and dramatic implications of the opinion became clear. At a time when sentencing had become driven in most states and the federal system by sentencing guidelines that were often calculated by judicial fact-finding, the follow-up decisions in Blakely v. Washington (542 U.S. 296 (2004)) and United States v. Booker (543 U.S. 220 (2005)) fundamentally changed the way sentencings operated. Booker ultimately struck down the mandatory nature of the federal sentencing guidelines and similar state systems. All along the way, Kennedy was in dissent, as were Rehnquist, O’Connor, and Breyer. The Apprendi line of cases, pressed most vigorously on the right by Scalia, provide an example of “outside-in” majority coalitions in which the most conservative and liberal members of the Court join forces against the moderates. Kennedy was sure to be on the losing side in such cases.

Here, Kennedy’s absence seems unlikely to have much impact, except to further strengthen the coalition in support of Apprendi’s principles. Based on Kavanaugh’s writings before joining the Supreme Court, it is a safe bet to say that he will reliably vote to continue the Scalia-driven revolution.

The Sixth Amendment and the Confrontation Clause

It might be an overstatement to describe the Crawford v. Washington line of cases as revolutionary (the palace guards have won a few recent battles), but not by much. (541 U.S. 36 (2004).) Here, unlike the Apprendi line of cases, Kennedy wasn’t entirely opposed from the start, but in the most important Crawford cases aside from Crawford itself, he has voted for the prosecution.

Kennedy signed on to Scalia’s opinion in Crawford, which overruled the longstanding reliability exception to the use of hearsay statements against an accused as set forth in Ohio v. Roberts. (448 U.S. 56 (1980).) Crawford held that out-of-court “testimonial” statements offered against an accused are barred by the Confrontation Clause unless the witness is unavailable and the accused had a prior opportunity to cross examine the witness. Notably, Kennedy did not join Rehnquist and O’Connor in their dissent from the overruling of Roberts (they concurred in the judgment by applying Roberts).

The big question in nearly every Crawford case is the definition of “testimonial.” And although Kennedy joined the Crawford majority, he was almost entirely a pro-government vote in the significant Crawford follow-up cases. He wrote the dissents in the 5–4 decisions in Bullcoming v. New Mexico (564 U.S. 647 (2011)) and Melendez-Diaz v. Massachusetts (557 U.S. 305 (2009)), which held that uses of laboratory reports offered by the prosecution against the accused without calling the lab analysts who conducted the testing violated the Confrontation Clause. He joined the majority ruling against the defendant in Michigan v. Bryant, which altered the Crawford analysis to make it more favorable to the prosecution. And in the last closely divided Crawford decision, Williams v. Illinois, he joined a four-justice plurality whose reasoning, if it were eventually fully adopted by the Court, would greatly diminish the Crawford protections against the use of out-of-court statements against the accused. (567 U.S. 50 (2012).)

Here, there is good reason to believe that Kennedy’s retirement will benefit defendants. First, short of overruling Crawford, it is hard to imagine any realistic way in which his replacement could be worse for defendants than he was. And second, there seems to be as much reason to think that Kavanaugh will follow in the pro-defendant footsteps of Scalia on confrontation rights as he will follow in the pro-prosecution footsteps of Roberts and Alito.

The Sixth Amendment and the Right to Counsel

In the companion cases Missouri v. Frye, 566 U.S. 134 (2012) and Lafler v. Cooper, 566 U.S. 156 (2012), Kennedy wrote the opinions and provided the necessary fifth votes siding with the four liberals in expanding the Sixth Amendment’s right to effective assistance of counsel to cover plea bargaining. In Lafler, the defense lawyer had provided his client patently incorrect legal information causing the client to reject a plea bargain and go to trial, which ended predictably poorly. In Frye, the defense lawyer failed to convey a favorable plea offer to his client who later pled to a worse deal. In both cases Kennedy found the attorneys’ performances fell well below professional standards and caused prejudice, thereby violating the Sixth Amendment.

The four conservatives dissented, with Scalia writing to rail against the “whole new field of constitutionalized criminal procedure: plea bargaining law.” In practical terms, the rhetoric was likely overheated as defendants rarely prevail on these claims (just as they rarely prevail on ineffective assistance of counsel claims generally). But the opinions did in fact expand Sixth Amendment protections beyond their previous contours, and Kennedy’s statement that the American criminal justice system “is for the most part a system of pleas, not a system of trials” gave the Supreme Court’s imprimatur to a fact long known and highly criticized by observers of the system.

The durability of Frye and Lafler almost surely depends on Roberts’ allegiance to stare decisis. It is hard to imagine Kavanaugh (or Gorsuch) taking a different approach than that of the dissenting conservative justices.

The Fourth Amendment

Kennedy’s votes on the Fourth Amendment were mostly pro-government, but in some of the more consequential cases his vote was not so pivotal. For instance, in Utah v. Strieff, 136 S.Ct. 2056 (2016) he was one of six to hold against the defendant and water down the exclusionary rule in the face of a clear Fourth Amendment violation. In 2013 Kennedy was part of the dissent in Florida v. Jardines, which held that police officers’ use of a drug-sniffing dog on the front porch of a home was a trespassory encroachment onto the curtilage constituting an unlawful search. (569 U.S. 1 (2013).) Along with Roberts, Alito, and Breyer, he would have held that the activity was neither a trespass nor an invasion of a reasonable expectation of privacy. He had earlier joined a similarly mixed group of four dissenters in Kyllo v. United States in opposing the Court’s ruling that the police use of thermal imaging to measure heat emanating from a home was an unlawful search. (533 U.S. 27 (2001.))

Although he ruled for defendants in two of the Court’s more recent high-profile cases involving new technologies, neither was a close vote. In Riley v. California (134 S. Ct. 2473 (2014)), a unanimous Court held that searches of cell phones require a warrant, and in United States v. Jones (565 U.S. 400 (2012)), a unanimous Court (though divided on the reasoning) held that law enforcement’s affixing of a GPS tracking device to a car and monitoring its movements was a search within the meaning of the Fourth Amendment. In Jones, Kennedy joined the more conservative reasoning that the decision should be grounded in the law of trespass.

In the most recent and significant Fourth Amendment case, Carpenter v. United States (138 S. Ct. 2206 (2018)), Kennedy authored a dissent to the five-justice majority opinion written by Roberts holding that individuals maintain a legitimate expectation of privacy in the cell-site location information stored by wireless phone carriers and that a warrant is therefore generally required before that information may be obtained by law enforcement. Carpenter was a major exception to the doctrine negating any Fourth Amendment interest individuals might have in records maintained by third parties, and Kennedy described the decision as a “stark departure” from the Court’s precedents that would put “criminal investigations at serious risk in serious cases.” Here again, Kennedy’s departure can only improve defendants’ future prospects.

Kennedy’s votes were critical in two particularly noteworthy cases restricting Fourth Amendment rights in which he provided the fifth vote for the government. In Maryland v. King, he authored the opinion allowing police to routinely collect DNA samples from arrestees, just as they may take fingerprints or mug shots, on the ground that it was justified to identify the person in custody. (569 U.S. 435 (2013).) The flimsy rationale was exposed as such by a scathing Scalia dissent which was joined by Ginsberg, Sotomayor, and Kagan. And in Hudson v. Michigan , he joined a 5-4 majority decision holding that the violation by the police of the knock-and-announce rule did not require suppression of all of the evidence subsequently found in the search of a home, continuing his reliable vote to erode the exclusionary rule. (547 U.S. 586 (2006).)

Although his was a pro-government vote in Hudson, at least one prominent commentator thinks his concurrence in the case showed how much worse things could have been for defendants. In it, Kennedy wrote the following: “the continued operation of the exclusionary rule, as settled and defined by our precedents, is not in doubt.” (Id. at 603 (Kennedy, J., concurring in part and concurring in the judgment).) Of course, anytime a justice writes separately to note that a doctrine is not in doubt, it probably is. And, indeed, the plurality opinion of Roberts, Scalia, Thomas, and Alito in Hudson, which discussed the “considerable” social costs of applying the exclusionary rule and reviewing the instances in which it did not apply, made it so. As Dean Erwin Chemerinksy wrote earlier this year about the exclusionary rule in describing the possible impact of Kavanaugh on the Court, “with Kennedy being replaced by Kavanaugh, there could be five votes to overturn a doctrine that conservatives have opposed for decades.” (Erwin Chemerinsky, What’s at Stake if Kavanaugh Is on the Supreme Court, A.B.A. J., Aug. 29, 2018.) If so, it would certainly qualify as perhaps the single biggest consequence of Kennedy’s retirement on matters of criminal justice.

The Fifth Amendment and Vagueness

Another highly consequential line of cases, particularly for federal defendants, is Johnson v. United States (135 S. Ct. 2551 (2015)) and its progeny. In Johnson, the Court struck down on vagueness grounds the so-called residual clause of the federal Armed Career Criminal Act. Although Kennedy joined in the judgment, he wrote separately to disagree with the vagueness holding. And he joined the four-member dissent in the follow-up case, Sessions v. Dimaya, which held that a similar clause defining “crime of violence” in the federal code was also void for vagueness. (138 S. Ct. 1204 (2018.)) His absence will not likely make much of an immediate difference in the cases to follow.

Habeas Corpus and “Enemy Combatants”

Although the “enemy combatant” cases involving detainees at Guantanamo Bay did not impact the day-to-day workings of the criminal justice system, they were some of the most important decisions in the Court’s history, setting limits on Executive Branch power and staking a claim for the Court’s authority to address where the limits lie. And Kennedy’s role was critical. In Hamdan v. Rumsfeld, Kennedy provided the necessary fifth vote for the Court’s holding that it had jurisdiction to determine the lawfulness of military commissions in Guantanamo Bay and that the commissions as constituted were unlawful. (548 U.S. 557 (2006).) And in Boumediene v. Bush, Kennedy wrote the majority opinion in the 5–4 decision holding that “enemy combatants” held at Guantanamo Bay were entitled to the protections of habeas corpus and that the Military Commissions Act unconstitutionally suspended the right. (553 U.S. 723 (2008).) It seems a good guess that Kavanaugh, having worked in the George W. Bush White House, would be more inclined toward the Executive on these issues. It is hard to know how Roberts may react in his new role and given the precedent of Hamdan and Boumediene.


In some of the most consequential criminal justice issues of his day, Kennedy was the single most pivotal member of the Supreme Court. In the areas of capital punishment and juvenile rights, defendants are likely to fare poorly in his absence. Just how poorly likely depends on Roberts and his allegiance to stare decisis, as there is little reason to think that Kavanaugh will view those capital and juvenile rights decisions as Kennedy did. In other areas, particularly Miranda, Apprendi, search and seizure, and Confrontation Clause cases, Kennedy’s departure will either make little difference to criminal defendants or improve their outlook.

Of course, guesses about the future votes of Kavanaugh and a post-Kennedy Roberts are just that. Justices have often surprised in their time on the Court. The Sphynx of Sacramento, as some have called Kennedy, was no different. Nothing in his past, or even in the first half of his 30-year tenure on the Court, would have foretold his later role as vigorous champion for death row inmates and juveniles against an equally vigorous conservative dissent.

Kennedy’s legacy is now in the hands of his former colleagues, and it may well be that his most lasting decision was his decision to retire when he did.


David Patton

David Patton is the executive director of the Federal Defenders of New York and an adjunct law professor at New York University.