Precedent, and Departing From It

It is used sparingly. Lord Reid's observation in Knuller, quoted in Austin, is the standing statement: the change of practice does not mean that whenever the House thinks a previous decision was wrong it should reverse it, and in the general interest of certainty there must be some very good reason before it does so.

The American practice

There is no text. Article III says nothing about precedent, and the Supreme Court's power to overrule its own decisions is inherent in the judicial power.

What restrains it is a set of considerations, of which the strongest is reliance: whether people have arranged their affairs on the faith of the decision. Others are whether the rule has proved unworkable, whether the reasoning has been eroded by later decisions, and whether the facts or the understanding of the facts have changed.

The example on this syllabus is Brown v. Board of Education of Topeka, 347 U.S. 483 (1954), which held that separate educational facilities are inherently unequal and that the doctrine of separate but equal has no place in public education, displacing Plessy v. Ferguson, 163 U.S. 537 (1896), after fifty eight years.

And the constitutional reason overruling matters more there. A statutory interpretation the Court gets wrong can be corrected by Congress; a constitutional interpretation can be corrected only by amendment, which is why the Court has treated the case for reconsidering a constitutional decision as stronger than for a statutory one.

Switzerland

The Federal Supreme Court is not bound by its own decisions in the common law sense, but it publishes leading decisions in the official BGE collection and departs from them only for compelling reasons, giving its reasons for the change when it does. A change of practice must be announced rather than smuggled in, and there is no doctrine of binding precedent for the lower courts comparable to article 141.

Comparison

Prospective overruling

A court that overrules a decision unsettles everything done on the faith of it. One answer is to declare the new rule but apply it only to the future.

The Indian Supreme Court adopted the technique in I.C. Golak Nath v. State of Punjab, AIR 1967 SC 1643: it held that Parliament had no power to amend Part III so as to take away or abridge fundamental rights, but applied the holding prospectively, so that the earlier amendments already made were left standing.

The device is criticised as legislative, since a court that announces a rule for the future and does not apply it to the parties before it is not deciding their case on the law. It is defended as the only way to correct an error without destroying settled transactions.

A worked example

A five judge bench of the Indian Supreme Court in 2004 decided a point one way. A two judge bench in 2026 thinks it wrong.

It cannot depart. It refers the matter, and the Chief Justice constitutes a bench of at least five, or of more if the earlier decision is to be reconsidered rather than distinguished.

In the United Kingdom, a differently constituted Supreme Court could hear the point and depart under the Practice Statement, subject to Austin's caution about reliance.

In the United States, the same nine judges would hear it, and reliance would be the central question.

The cases

Facts. Austin v. Mayor and Burgesses of the London Borough of Southwark, [2010] UKSC 28, concerned succession to a secure tenancy and the status of a tolerated trespasser, which earlier decisions of the House of Lords had settled in a way the Court doubted; the Court was asked to depart from them.

Held, by Lord Hope with the Court. The power to depart exists in the Supreme Court. The Practice Statement of 26 July 1966 was part of the established jurisprudence relating to the conduct of appeals in the House of Lords which was transferred to the Supreme Court by section 40 of the Constitutional Reform Act 2005, and it was unnecessary to re-issue it. But the Court declined to depart: the law had long been regarded as settled, had been acted on in many tens of thousands of cases, and the effects of a retrospective reversal on social landlords were incalculable.

Why it matters here. It states the modern British rule and shows what actually decides these cases, which is reliance rather than correctness.

Facts. Brown v. Board of Education of Topeka, 347 U.S. 483 (1954), consolidated challenges by Black schoolchildren to state laws requiring or permitting racial segregation in public schools.

Held, by Warren CJ for a unanimous Court. Separate educational facilities are inherently unequal, and segregation in public education deprives the plaintiffs of the equal protection of the laws; the doctrine of separate but equal has no place in public education.

Why it matters here. It is the standing demonstration that an entrenched constitutional decision can be displaced by the same court that made it, and that where amendment is unavailable the court's willingness to reconsider is the only route.

Facts. I.C. Golak Nath v. State of Punjab, AIR 1967 SC 1643, was a challenge to constitutional amendments placing land reform legislation beyond challenge, and required the Court to consider its own decisions in Sankari Prasad and Sajjan Singh.

Held, by eleven judges, six to five. Parliament has no power to amend Part III so as to take away or abridge a fundamental right, an amendment being law within article 13(2); but the decision would operate prospectively, so that the amendments already made would stand.

Why it matters here. It shows both Indian devices at once: a larger bench displacing smaller ones, and prospective overruling used to contain the consequences.

Distinctions

What beginners get wrong

"Article 141 binds the Supreme Court." It binds all courts within the territory of India, and the settled construction is that it does not bind the Supreme Court itself.

"Obiter is worthless." It is not binding, but obiter of the Supreme Court is treated as of high persuasive value and lower courts follow it in practice.

"After 1966 the Court of Appeal can depart from its own decisions." The Practice Statement was made by the House of Lords for itself.

"The Practice Statement lapsed in 2009." Austin holds it did not; it transferred with the jurisdiction under section 40 of the Constitutional Reform Act 2005.

Limits and criticism

Bench strength is a crude measure. A point may be settled by numbers rather than by reasons, and constituting a larger bench is in the gift of the Chief Justice.

Prospective overruling asks a court to legislate, and to decide the case before it by a rule it has just declared wrong.

Reliance protects the status quo, which means the longer an error stands the safer it becomes, and Austin is candid about that.

Departure without acknowledgment is the real problem in every system: a decision quietly not followed leaves the law less certain than one openly overruled.

Quick revision

1. Article 141: the law declared by the Supreme Court is binding on all courts within the territory of India.
2. It states the vertical rule; the Supreme Court is not bound by itself, and bench strength governs.
3. Article 145(3): minimum of five judges for a substantial question of constitutional interpretation.
4. Practice Statement (Judicial Precedent), [1966] 1 WLR 1234: normally binding, but departure where it appears right.
5. Austin, [2010] UKSC 28: the Statement transferred under section 40, Constitutional Reform Act 2005; reliance defeated departure.
6. Brown, 347 U.S. 483 (1954), displacing Plessy, 163 U.S. 537 (1896).
7. Golak Nath, AIR 1967 SC 1643: prospective overruling.

Test yourself

1. What does article 141 settle and what does it leave open? It settles that the law declared by the Supreme Court binds all courts within the territory of India, which is the vertical rule, and it settles that what binds is the law declared, the ratio, and not every observation. It leaves open whether the Supreme Court is bound by its own decisions. The answer worked out in practice is that it is not, but that a bench may not depart from a decision of a bench of equal or greater strength and must refer the question, so that departure is achieved by constituting a larger bench.

2. State the Practice Statement and its position today. Issued by the House of Lords on 26 July 1966, reported as Practice Statement (Judicial Precedent) [1966] 1 WLR 1234, it stated that former decisions of the House would still be treated as normally binding but that it would depart from a previous decision when it appeared right to do so. Austin v. Mayor and Burgesses of the London Borough of Southwark, [2010] UKSC 28, holds that the Supreme Court need not re-issue it, since it formed part of the established jurisprudence on the conduct of appeals transferred by section 40 of the Constitutional Reform Act 2005, and has the same effect in the Supreme Court.

3. Why does overruling matter more for an American constitutional decision than for a statutory one? Because the routes of correction differ. A statutory interpretation the Court gets wrong can be corrected by Congress passing another statute; a constitutional interpretation can be corrected only by an Article V amendment, which requires two thirds of both Houses or a convention and ratification by three quarters of the States. Where the political correction is practically unavailable, the Court's own willingness to reconsider is the only route, which is why Brown had to displace Plessy rather than wait for an amendment.

4. What is prospective overruling and what is wrong with it? It is the declaration of a new rule with effect only for the future, the transactions already completed on the faith of the old rule being left undisturbed; the Indian Supreme Court adopted it in I.C. Golak Nath v. State of Punjab, AIR 1967 SC 1643. The objection is that a court exists to decide the case before it according to law, and a court that announces the law is one thing and then decides the case by a rule it has just held wrong is making a legislative choice about commencement rather than adjudicating.