Stare Decisis Explained, and the Factors Justices Weigh Before Overruling

There are two ways to ruin a legal system with precedent. One is to treat every old decision as a mistake waiting to be reversed, so that nothing is ever settled. The other is to treat every old decision as scripture, so that nothing is ever corrected. A republic needs a rule that avoids both, and the rule has a Latin name that sounds like a spell.

Stare decisis. "To stand by things decided." It is the doctrine that courts should generally follow earlier decisions, and the argument over its limits is among the most heated in American law. Let us take it apart, piece by piece, and see where it holds and where it gives.

What stare decisis is, and is not

Start with the two kinds, because people mix them up.

Vertical stare decisis is the rule that lower courts must follow the decisions of higher courts. A federal district judge in Ohio is bound by the Supreme Court and by the Sixth Circuit. This is not optional and not controversial. It is the structure of the judiciary. A district judge who decides that a Supreme Court precedent is wrong is not entitled to ignore it. That judge may say so, in a careful opinion, in the hope the higher court will reconsider. But the judge must follow it.

Horizontal stare decisis is the question of whether a court should follow its own prior decisions. That is where the arguments live. The Supreme Court is not bound by its earlier decisions in the way a lower court is, because no higher court can say otherwise. It can overrule itself. The question is when it should.

Two more distinctions help. A decision interpreting a statute is on firmer ground than a decision interpreting the Constitution, because Congress can fix a statutory mistake by passing a new law, whereas a constitutional mistake can be fixed only by amendment, which is hard, or by the Court. The Court has said so for a long time, and has said that statutory precedents have special force for that reason. And "precedent" means the holding, the rule needed to decide the case, not every sentence in the opinion. The rest is dicta, and a court need not follow it.

Blackstone: judges declare, they do not make

The theory behind the doctrine comes from the English common law. Sir William Blackstone, in his Commentaries on the Laws of England in the 1760s, described judges as the "living oracles" of the law, men sworn to determine not according to their own private judgment but according to the known laws and customs of the land. They do not pronounce a new law but maintain and expound the old one.

Blackstone also supplied the escape hatch. A precedent should be followed, he said, unless it is "flatly absurd or unjust." Even then, he added, the proper way to understand the case is that the earlier judge was mistaken about the law, not that the law changed. The law was always what it was, and the earlier court got it wrong.

That is the declaratory theory, and it matters for a conservative reading of the judge's job. If judges declare the law rather than make it, then correcting an erroneous precedent is not an act of power. It is a return to what the law always was. Alexander Hamilton said in Federalist 78 that, "to avoid an arbitrary discretion in the courts, it is indispensable that they should be bound down by strict rules and precedents." The same Hamilton also wrote that courts must follow the Constitution over a statute, and by extension over an earlier decision that contradicts it.

So the Founders' view had two sides. Courts should be bound by precedent, because arbitrary discretion is a danger. And courts should obey the Constitution above all, because the Constitution is the people's law and precedent is only the judges' gloss on it.

Why precedent has value

Before arguing about when to overrule, state the case for sticking. It is strong.

1. Stability. A rule that changes with every new majority is not law but preference. People need to know what the law is before they act.
2. Equal treatment. Like cases should be decided alike. A court that decides this litigant's case differently from last year's, for no reason except that the membership changed, is being unfair to both.
3. Reliance. Contracts, property, businesses, marriages and mortgages are made on the assumption that the law will stay the same. Pulling the rug out can cost people real money and real security.
4. Humility. A judge who thinks every past decision is subject to his own reading is a judge who thinks he is smarter than all of his predecessors. That is rarely a good bet.
5. Legitimacy. The Court has no army or budget. It has the habit of being obeyed, and that habit depends on a perception that the Court follows rules, not the tides.

Those are real, and any conservative who says "precedent is nothing" will discover he believes in it the first morning a precedent he likes is in danger. Everybody is a precedent skeptic about the other side's cases and a precedent fundamentalist about his own.

Payne v. Tennessee: not an inexorable command

In Payne v. Tennessee (1991), the Court overruled two recent decisions, Booth v. Maryland (1987) and South Carolina v. Gathers (1989), that had barred prosecutors from presenting victim-impact evidence in capital sentencing. Chief Justice Rehnquist wrote for the Court that stare decisis "is not an inexorable command." It is a principle of policy, he said, and not a mechanical formula of adherence to the latest decision. He added that the doctrine is at its strongest in cases involving property and contract rights, where reliance is high, and weaker in cases involving procedural and evidentiary rules.

Payne shows two things. The Court can overrule even a recent decision, and it will do so when the earlier decision was badly reasoned and had not generated real reliance. Booth and Gathers were only a few years old and had been decided by narrow majorities over vigorous dissents. Justice Marshall's dissent in Payne accused the majority of overruling precedent only because the membership had changed, and that charge gave the argument its sharpest edge. The answer to it is the same one that applies to every case: a bad decision does not become good by being young, and a good decision does not become safe by being old.

Casey's reliance argument

A year later, in Planned Parenthood v. Casey (1992), three Justices used stare decisis to defend Roe. Their joint opinion argued that overruling would harm the Court's legitimacy, because the public would see it as caving to political pressure, and that reliance supported leaving the right alone: that people had organized intimate relationships and made choices that define their views of themselves and their place in society, in reliance on the availability of abortion.

Take that argument seriously. It is the best statement of the pro-precedent position in recent history. It says that a long-standing rule creates expectations, and that to remove it is to harm those who relied on it. It says the Court should not overrule under pressure, since to do so would make every future decision look political.

Then notice the problem. The legitimacy argument, taken seriously, would make any wrongly decided case permanent as soon as it became controversial, because to overrule it would then look political. The more controversial the error, the less it could be corrected. That is the reverse of what a constitutional court should do. Legitimacy, in the end, comes from getting the law right, and from willingness to say when a prior court did not.

Kavanaugh's three questions in Ramos

In Ramos v. Louisiana (2020), the Court held that the Sixth Amendment requires a unanimous jury to convict of a serious crime, overruling Apodaca v. Oregon (1972). Justice Kavanaugh wrote a concurrence in part that is the most useful single statement of how to think about overruling. He asked three questions:

1. Is the prior decision not just wrong, but grievously or egregiously wrong?
2. Has the prior decision caused significant negative jurisprudential or real-world consequences?
3. Would overruling the prior decision unduly upset reliance interests?

He then applied them. Apodaca was egregiously wrong, he said, because its reasoning was gravely mistaken. It had real consequences in the form of non-unanimous verdicts that, as the Court's opinion described, had roots in the racial politics of Louisiana and Oregon. And reliance interests were weak, because the rule affected criminal procedure and not property or contracts, and the States could manage the transition. So it should go.

Three questions, applied honestly, are a good discipline. Notice that they require an answer to each. A judge cannot overrule just because he thinks the earlier decision was wrong, and cannot refuse to overrule just because it is old.

The factors in Janus and Dobbs

The Court has also set out lists of factors. In Janus v. AFSCME (2018), overruling Abood v. Detroit Board of Education (1977), Justice Alito identified several considerations: the quality of the earlier decision's reasoning, the workability of the rule it established, its consistency with other decisions, developments since it was decided, and the reliance interests it had created.

In Dobbs v. Jackson Women's Health Organization (2022), the Court gave a list of five: the nature of the error, the quality of the reasoning, workability, the effect on other areas of the law, and reliance interests. You can see that these overlap, and that they work as an outline of a good argument rather than a mechanical test. A court that wants to overrule must show, on each factor, why the decision cannot stand. That is a heavier burden than a bumper sticker, which explains why nobody has yet printed one.

Compare the lists and a pattern emerges. All of them ask three things, in different words: was it wrong, has it worked, and who will be hurt if it goes?

Three misunderstandings

"Overruling is activism." It is not. A court that overrules a decision it thinks erroneous is applying the Constitution as it understands it. Activism is not about whether a decision is old or new. It is about whether the judge is following the text or his own preference.

"Overruling is conservative, and following precedent is liberal." This is just wrong. Brown v. Board of Education overruled Plessy. Gideon overruled Betts. Barnette overruled Gobitis. Those are among the finest moments in American law, and they were all overrulings. Conservatives are no more opposed to correcting errors than anyone else.

"A mistake that has been repeated is no longer a mistake." Time lends a mistake no authority. Errors can calcify, and ignoring them is a choice. But time does matter to the question of what reliance has grown up around them.

The strongest objection

The strongest objection is that all these factors are subjective. "Egregiously wrong" and "poorly reasoned" are labels for a judge's view, and a Court with five votes can decide that any inconvenient precedent is egregiously wrong. The factors, on this view, are a cover for a result.

That risk exists, and it should be said plainly. The way to guard against it is not to abandon the factors but to insist on discipline in applying them.

• Demand a showing, from text and history, that the earlier decision was wrong, rather than a mere preference.
• Demand evidence that the rule has not worked in practice, rather than just a feeling.
• Demand a real accounting of reliance, and a transition rule where the reliance is genuine.
• Demand that the Court say what it is doing, rather than narrowing a precedent into nothing without ever admitting it.

A Court that meets those demands is accountable to the text. A Court that does not is no better than the Court it replaced.

What it means for a family

A family cares about precedent in a concrete way. When you sign a mortgage, start a business or write a will, you rely on the law staying put. That is stare decisis at work, and it is why the Court gives property and contract the greatest weight. And when the law is wrong in a way that harms people, a family cares about correction. A rule that locks innocent people into non-unanimous convictions, or compels workers to support a union they did not join, does not become right by repetition.

So hold both thoughts. Respect precedent, because stability serves the ordinary citizen first. And insist that the Court say what the Constitution means, because the Constitution, not the last opinion, is the law.

Read more in From Plessy to Chevron, a Timeline of Precedents the Court Was Right to Overrule, which gives the dated list, Dobbs v. Jackson Read Closely, From Alito's History to the Joint Dissent for the factors at work, and Ramos v. Louisiana and the Jim Crow Roots of Split-Jury Convictions for Kavanaugh's three questions in action. The debate over Roe's staying power is in Roe v. Wade Was Weak Law Even to Scholars Who Liked the Result, and the Founders' view of the judge's role is in Federalist 78 Called Courts the Least Dangerous Branch. Was Hamilton Right?. See also Why Originalism Is the Only Method That Binds the Judge as Well as the Citizen. The Court's own pages are linked from the Courts section.

The case in brief

1. Vertical stare decisis binds lower courts; horizontal stare decisis is the Supreme Court's policy toward its own decisions, and it is a policy, not an inexorable command.
2. Blackstone's rule is to follow precedent unless it is flatly absurd or unjust, and to treat an overruling as a correction of error, not a change in the law.
3. The Court's working tests, Kavanaugh's three questions in Ramos and the factors in Janus and Dobbs, ask whether the decision was wrong, whether it has worked, and who relied on it.
4. Precedent should be strongest where property and contract rights depend on it and weakest where a constitutional error harms liberty.
5. The discipline is to show the error, the damage and the reliance, and to say plainly what the Court is doing.

What you can do: read the first two pages of Kavanaugh's concurrence in Ramos on the Supreme Court's website and apply his three questions to a precedent you disagree with.