Stare Decisis and the Rules for When the Court Reverses Itself

Elsie Parrish cleaned rooms at the Cascadian Hotel in Wenatchee, Washington, a town in the apple country east of the Cascades. Washington had a minimum wage law for women and minors, and when she left the job she calculated that the hotel had paid her less than it required. She sued for the difference, a little over two hundred dollars. The hotel's defense was simple and, at the time, persuasive: the law was unconstitutional. In 1923, in Adkins v. Children's Hospital, the Supreme Court had struck down a minimum wage law for women in the District of Columbia as a violation of the freedom of contract. As recently as 1936, the Court had applied that precedent to strike down New York's minimum wage law. Parrish had every reason to lose. On March 29, 1937, she won, five to four, and Adkins was gone. Her case is one of the clearest examples of the Court reversing itself, and the way it happened tells you a lot about the rule that is supposed to stop that from happening.

In short: Stare decisis, Latin for "to stand by things decided," is the principle that courts should generally follow their own past decisions. The Supreme Court can and does overrule itself, and over time it has developed factors to decide when: how badly reasoned the old decision was, whether it has proved workable, how much people have relied on it, and whether later law or facts have eroded it. Both conservatives and progressives praise the principle when it protects decisions they like and set it aside when it does not.

Why courts bind themselves at all

The idea behind stare decisis is older than the country. Courts follow their own precedents for several practical reasons. It makes the law predictable, so people and businesses can plan. It treats like cases alike, which is a basic form of fairness. It saves judges from having to relitigate every question from scratch. And it restrains them, since a judge who must follow precedent cannot simply impose personal views.

Justice Louis Brandeis put the core of it in a 1932 dissent, in a case called Burnet v. Coronado Oil and Gas Company: in most matters, he wrote, "it is more important that the applicable rule of law be settled than that it be settled right." That line is quoted often because it captures the trade-off. Sometimes a rule that is slightly wrong but stable is better than a rule that keeps changing.

But Brandeis went on, in the same opinion, to explain why constitutional cases are different. When the Court gets an ordinary statute wrong, Congress can fix it by passing a new law. When the Court gets the Constitution wrong, the only fix short of a constitutional amendment is for the Court itself to change its mind. So in constitutional cases, he argued, the Court should be more willing to correct its errors. Every modern debate about overruling precedent echoes that tension.

The switch, and what it was not

Back to Elsie Parrish. Her victory is often told as part of a dramatic political story. In February 1937, President Franklin Roosevelt, frustrated that the Court kept striking down New Deal laws, proposed adding up to six new justices, a plan quickly labelled court packing. Weeks later, Justice Owen Roberts, who had voted with the majority to strike down New York's minimum wage law the year before, voted with the majority to uphold Washington's. Journalists called it "the switch in time that saved nine."

The story is better than the truth, though the truth is more interesting. Records later showed that the justices had voted on Parrish's case in conference in December 1936, before Roosevelt announced his plan. Roberts had already changed his position. The decision was announced in March because Justice Harlan Stone had been ill and the Court waited for him. Roberts also later explained that he had voted as he did in the New York case partly because of how that case had been argued, since New York had not directly asked the Court to overrule Adkins.

Chief Justice Charles Evans Hughes's majority opinion in West Coast Hotel Co. v. Parrish did not tiptoe. It said the Constitution does not speak of freedom of contract, that liberty is subject to reasonable regulation in the interests of the community, and that the economic conditions of the Depression had made plain the cost of letting employers pay women wages too low to live on. Adkins, the opinion concluded, should be and was overruled.

The factors justices name

There is no statute listing when the Court should overrule itself. Instead, justices have built up a set of considerations over many decades. Different opinions state them differently, but most lists include these.

The quality of the reasoning. Was the old decision well reasoned, or did it rest on an analysis that was weak from the start? This is the factor most likely to be shaped by a justice's broader judicial philosophy, which is why it is the most contested.

Workability. Has the rule given lower courts and the public a clear standard, or has it produced confusion and inconsistent results? A rule that judges cannot apply consistently has less claim to be preserved.

Consistency with related decisions. Has the old precedent become an outlier, out of step with the Court's other cases? Doctrines sometimes grow in one direction and leave earlier cases stranded.

Changed facts or understanding. Have the facts on which the decision rested changed, or come to be seen differently? Brown v. Board of Education, which in 1954 overruled the separate-but-equal doctrine of Plessy v. Ferguson for public schools, relied partly on evidence about the effects of segregation on children.

Reliance. Have people, businesses or governments organized their lives and affairs around the old rule? Reliance weighs most heavily in areas like property and contract, where people make concrete plans based on settled law.

The Court's 1992 opinion in Planned Parenthood v. Casey set out a version of these factors and used them to reaffirm the core of Roe v. Wade. Thirty years later, the majority in Dobbs v. Jackson Women's Health Organization used a similar list and reached the opposite conclusion, saying that the reasoning in Roe was exceptionally weak and that the reliance interests at stake were not the concrete, economic kind the doctrine traditionally protects. Same tools, different hands, different result.

Statutes get extra protection

There is one area where the Court has said stare decisis carries special weight: decisions interpreting statutes, as opposed to the Constitution. The logic follows from Brandeis. If Congress disagrees with how the Court read a law, it can amend the law. If Congress leaves the reading in place for years, the Court treats that as a reason not to disturb it.

The best-known modern statement of this came in a 2015 case, Kimble v. Marvel Entertainment, about royalties on a toy that let children shoot foam string from their wrists like Spider-Man. The question was whether a patent holder could keep collecting royalties after the patent expired; a 1964 precedent said no. Justice Elena Kagan, writing for the majority, declined to overrule it, noting that Congress was free to change the rule. Her opinion was laced with references to the comic book, including the line that with great power there must also come great responsibility. It is one of the rare Supreme Court opinions that is both an important statement of doctrine and genuinely fun to read.

Even here, though, the protection is not absolute. In 2024, in Loper Bright Enterprises v. Raimondo, the Court overruled the 1984 Chevron decision, a doctrine about how courts should treat agency interpretations of statutes. The majority reasoned that Chevron was a judge-made rule of interpretation rather than a reading of any particular statute, so the extra weight for statutory precedents did not apply in the same way.

Everyone's principle, when convenient

Here is where honesty requires discomfort on all sides.

Progressives celebrate many overrulings. Brown overruled Plessy. Lawrence v. Texas in 2003 overruled Bowers v. Hardwick, a 1986 decision that had upheld criminal laws against same-sex intimacy. West Coast Hotel itself overruled Adkins. Gideon v. Wainwright in 1963 overruled an earlier decision that states need not provide lawyers to poor defendants. In each case, the precedent being overruled had been in place for years, and in each case people relied on it, in some sense.

Conservatives celebrate a different list. Citizens United v. FEC in 2010 overruled parts of earlier campaign finance decisions. Janus v. AFSCME in 2018 overruled a 1977 precedent allowing public-sector unions to collect fees from nonmembers. Dobbs overruled Roe and Casey. Loper Bright overruled Chevron.

Each side tends to say that its overrulings corrected grave errors, while the other side's overrulings show a lack of respect for precedent. That is not quite hypocrisy; it reflects the fact that stare decisis is not, and has never been, an absolute command. The Court itself has said that it is a principle of policy, not an inexorable command. The real disagreement is almost always about the underlying question, whether the old decision was wrong, rather than about stare decisis itself.

A better way to argue about it

Here is the argument this history supports. Progressives should stop treating stare decisis as a shield that ought to protect only their precedents, and start making a consistent case about what the factors should mean.

That means taking reliance seriously in its human form, not just its commercial form. When people plan families, careers and medical decisions around a right the Court has recognized for decades, that is reliance as real as a contract. It means treating workability as a question about whether ordinary people and lower courts can actually follow the rule, not as a pretext. It means accepting that some precedents progressives dislike, from cases on campaign finance to cases on arbitration, are entitled to the same analysis, and making the case against them on the merits.

It also means remembering Elsie Parrish. The point of overruling Adkins was that a rule protecting the freedom of contract had, in practice, protected employers' freedom to pay wages no one could live on. The Court did not reverse itself because politics demanded it. It reversed itself because the rule had failed the people it was supposed to govern. That is the strongest possible case for overruling, and progressives can make it without apology when it fits.

Only the Supreme Court can overrule the Supreme Court

There are two kinds of stare decisis, and the difference matters for anyone trying to understand why lower courts sometimes reach results that seem out of date.

The first, sometimes called horizontal stare decisis, is a court's practice of following its own earlier decisions. That is what this post has mostly been about, and it is a strong presumption rather than an absolute rule. The Supreme Court can overrule itself, and so can a federal court of appeals sitting with all of its active judges, which is called sitting en banc.

The second, vertical stare decisis, is much stricter. Lower courts must follow the decisions of higher courts, full stop. A federal district judge in Ohio must follow the Sixth Circuit and the Supreme Court, even if the judge believes those precedents are wrong. The Supreme Court has been firm about this. In a 1989 case, Rodriguez de Quijas v. Shearson/American Express, it said that if one of its precedents has direct application, lower courts should follow it even if later decisions seem to have undermined its reasoning, and leave to the Supreme Court the prerogative of overruling its own decisions.

That rule creates a strange but important dynamic. Lower court judges who disagree with a Supreme Court precedent often follow it while writing separately to say why they think it should be reconsidered. Those opinions are a signal, read closely by lawyers and by the justices themselves, that a precedent may be ripe for challenge. Litigants then bring cases designed to put the question squarely before the Court, as Mississippi did with its fifteen-week abortion law and as a group of herring fishermen did with Chevron.

It also means that the path to overruling a precedent runs through careful case selection over years. Movements that understand this, and that build cases patiently, tend to be the ones whose views eventually prevail.

What you can do

Whether a precedent survives is decided by nine people, but the arguments about it play out in public, and so do the confirmations of those nine.

1. Read what the Court actually decides. The court watch page lists Supreme Court opinions this term with the Court's own summaries, along with the emergency docket and Senate votes on judges. When the Court overrules a precedent, the summary will usually say so.
2. Follow the debate. The courts topic page gathers reporting on the Court's direction, including which precedents are under challenge.
3. Check the wage floor that Parrish made possible. Our state minimum wage table shows every state's current minimum, a law that exists in its modern form because of that 1937 case.
4. Ask your senators about judges. Senators confirm federal judges. Use the representatives lookup to find their phone numbers and latest votes, and ask how they weigh a nominee's views on precedent.

Elsie Parrish never set out to change constitutional law. She wanted the wages the law said she was owed. That is often how the biggest reversals begin.