Ouster Clauses: India's Unhappy Borrow from Anisminic

The ouster clause is one of those doctrinal objects that survives because every constitutional system needs some way of saying that, in some matters, a final administrative determination must actually be final. Anisminic gave the United Kingdom a doctrine for it. The American administrative-preclusion line gave the United States its own structured version. India borrowed both, but kept only the surface — the assertion that the constitutional courts can always look behind a finality clause — and discarded the underlying distinction between jurisdictional error and error within jurisdiction that the borrowed doctrine was built to enforce. The result is a finality clause that neither finalises nor disciplines review.

What Anisminic actually decided, and what it did not

The 1969 decision in Anisminic Ltd v Foreign Compensation Commission is, in the United Kingdom, the case that defined modern administrative-law doctrine on ouster clauses.¹ The Foreign Compensation Commission had been set up to determine claims arising from the Suez nationalisations. Section 4(4) of the Foreign Compensation Act 1950 provided that a determination of the Commission 'shall not be called in question in any court of law'. The wording was as absolute as a Parliament could draft. The House of Lords, by a 3-2 majority, held that the clause did not bar judicial review of a determination that was, in law, a nullity — that an error of law going to jurisdiction was not a determination at all, and so could not enjoy the protection of an ouster clause whose subject was 'determinations'.

The doctrinal move in Anisminic was specific. It expanded the category of jurisdictional error to cover almost any error of law made by a tribunal — failing to consider relevant matters, taking irrelevant matters into account, asking the wrong question, applying a wrong test. After Anisminic, the distinction between jurisdictional and non-jurisdictional error did not disappear; it was preserved as a doctrinal trellis on which the wider review was hung. The clause was respected as far as it could go; review followed because the cases in question were treated as falling outside the scope the clause could protect. The structural form of the doctrine survived.

The decision had to do work that the Indian doctrine has not asked of it. In the United Kingdom, Parliament is sovereign, and the courts cannot strike down an ouster clause as unconstitutional. Anisminic had to develop a construction that respected the clause on its terms and produced review on the merits. The form of the doctrine is shaped by that structural constraint. The Indian system, with Articles 32 and 226 of the Constitution in place, does not face the same constraint. The Indian courts have an easier exit. They can — and have — held that judicial review is itself part of the basic structure of the Constitution, and that no statutory finality clause can take it away.² If the constitutional shortcut is available, why retain the doctrinal apparatus of Anisminic?

The answer should be: because the apparatus is what disciplines the review. The shortcut is what abandons it.

The Indian inheritance: the form without the distinction

The standard Indian treatment of finality clauses tracks Anisminic on the surface. The administrative-law textbook formulation, drawn from a long line of decisions, runs roughly as follows. A finality clause does not bar review under Articles 32 and 226 of the Constitution. The clause does not protect action that is ultra vires the statute. The clause does not protect action that is procedurally ultra vires (which is treated as jurisdictional error). The clause does not protect action where the authority has decided a jurisdictional fact wrongly. The clause does not protect action that violates natural justice. The clause does not protect action that proceeds on a manifest error of law apparent on the face of the record.⁶

Each of these propositions has a doctrinal pedigree. Each is defended in the case law. Pabbojan Tea Co Ltd v Deputy Commissioner (1968) held that a finality clause does not bar review where the authority has violated mandatory procedure. Durga Shankar v Raghuraj (1954) held that Section 105 of the Representation of the People Act, which made every order of the Election Tribunal final, did not bar Article 32 review.⁴ Dhulabhai v State of Madhya Pradesh (1969) laid down a structured set of propositions on when the jurisdiction of civil courts is excluded.³ The form is there. The judicial vocabulary is there. The cases produce judgments that read like Anisminic's cousins.

What is missing is the discipline. The Indian treatment has not maintained the jurisdictional/non-jurisdictional distinction with anything like the rigour that the post-Anisminic United Kingdom maintained it for the two decades after the decision. The Indian Court has, instead, allowed the category of jurisdictional error to collapse into the broader category of 'patent error of law', and the category of 'patent error of law' to extend, at the Court's discretion, to any error of law that the bench in front of the case finds sufficiently grave to warrant intervention. That is not a doctrine. That is a discretion.

The administrative-law treatise expresses the move with unusual candour. There is, the text notes, 'now a tendency to abandon the distinction between jurisdictional facts and other facts which go to the merits. The reason is that there is no logical test to draw such a distinction. The courts tend to treat all facts as non-jurisdictional… as in the USA.' The textual aside is doing a lot of work. The distinction has not been abandoned; it has been dissolved into the merits without being said to be abandoned. The American analogy is invoked, but the American doctrine has not been imported. What has been imported is the American result — review of administrative determinations on the merits — without the disciplining categories the American doctrine uses to limit that review.

The American comparison, properly stated

It is worth being precise about the American doctrine the Indian textbooks gesture at. The United States does not have a unified doctrine of jurisdictional review. It has, instead, a body of administrative-preclusion law that begins with Marbury v Madison's assertion that 'it is emphatically the province and duty of the judicial department to say what the law is', and that has been elaborated through cases like Bowen v Michigan Academy of Family Physicians (1986) and Reno v Catholic Social Services (1993) into a structured presumption of judicial review of agency action.¹¹ The presumption can be displaced, but only by clear and convincing statutory evidence that Congress intended to preclude review. When the presumption is not displaced, review proceeds under the Administrative Procedure Act, which itself sets out the grounds — arbitrary and capricious, contrary to law, unsupported by substantial evidence, in excess of statutory authority.

What the American doctrine does, in other words, is identify both when review is available and what review consists of. Both halves of the doctrine are necessary. Without the first, statutory preclusion is meaningless. Without the second, review becomes whatever the reviewing court wants it to be. The American system has solved the problem by structuring both halves.

The Indian doctrine does the first half well enough. It has said, repeatedly and clearly, that constitutional review under Articles 32, 226, and 136 cannot be ousted by a finality clause.⁴ The constitutional point is not in doubt. What the Indian doctrine has not done is the second half. It has not identified, with any consistency, what review consists of when the constitutional courts intervene through their writ jurisdiction. The Tata Cellular trichotomy of illegality, irrationality, and procedural impropriety is the closest the doctrine comes to a structured account.⁷ It is borrowed openly from the GCHQ decision in the United Kingdom. But the categories are then applied with such elasticity that they do not constrain the analysis.

Wednesbury, proportionality, and the dissolution of structure

Prateek Jalan and Ritin Rai, in their Oxford Handbook chapter on the review of administrative action, trace exactly this dissolution.¹⁴ The Tata Cellular trichotomy is supposed to limit the review to particular grounds. The grounds — particularly 'irrationality, namely Wednesbury unreasonableness' — have not been kept distinct from broader Article 14 review based on 'manifest arbitrariness' or 'non-arbitrariness'. The terms have been used interchangeably. Mansukhlal Vithaldas Chauhan v State of Gujarat (1997) restates the appeal/review distinction with admirable clarity: the High Court does not sit as a court of appeal, but reviews the manner in which the decision was made, asking whether the authority exceeded its powers, committed an error of law, breached natural justice, reached a decision no reasonable tribunal would have reached, or abused its powers.⁹ That formulation, drawn directly from the post-GCHQ English vocabulary, is the structured doctrine the system needs. It is invoked. It is also routinely ignored. Shrilekha Vidyarthi v State of Uttar Pradesh (1991) admits the difficulty: 'The meaning and true import of arbitrariness is more easily visualized than precisely stated or defined. The question, whether an impugned act is arbitrary or not, is ultimately to be answered on the facts and in the circumstances of a given case.' That is not a doctrine. That is the disclaimer that no doctrine exists.

Compounding this, the Court has been unable to decide whether the doctrine of proportionality applies to administrative review, or whether Wednesbury unreasonableness still holds the field, or whether both apply simultaneously. McDowell (1996) calls the applicability of proportionality 'yet a debatable issue'.⁸ Hazarilal (2008) says that 'Wednesbury principle of unreasonableness has been replaced by the doctrine of proportionality'. K Shyam Kumar (2010) says 'it is not safe to conclude that the principle of Wednesbury unreasonableness has been replaced by the doctrine of proportionality' and that both apply. Special Reference No 1 of 2012 lists an entire catalogue of standards — fair, reasonable, non-discriminatory, transparent, non-capricious, unbiased — and treats them as collectively comprising Article 14.¹⁵ The constitutional bench in Subramanian Swamy v CBI (2014) confined Article 14 review of legislation to classification and excessive delegation, while leaving the wider grounds available against administrative action.⁵ The grounds available against administrative action have not been narrowed to anything that can be applied predictably.

This matters for ouster clauses because the only thing that limits the practical reach of a finality clause is the doctrine of review. If the doctrine of review is open-ended — if 'arbitrariness' can mean anything that a court finds objectionable on the facts before it — then a finality clause limits nothing. The clause does not even discipline the bench into having to articulate the jurisdictional ground on which it intervenes. The bench can intervene under 'arbitrariness' and call it review.

Jurisdictional facts and the Raza Textiles formulation

The Indian doctrine has had its Anisminic moment. Raza Textiles Ltd v Income Tax Officer (1973) holds that 'no authority, much less a quasi-judicial authority, can confer jurisdiction on itself by deciding a jurisdictional fact wrongly. The question whether the jurisdictional fact has been rightly decided or not is a question that is open for examination by the High Court in an application for writ of certiorari.'¹⁰ That is the Indian articulation of the jurisdictional-fact doctrine. It is a strong articulation. The High Court reviews the existence of the jurisdictional fact de novo, on the evidence, as if it were sitting in appeal on that limited question.

What has happened to Raza Textiles in the decades since? It survives as citation. It does not survive as doctrine. The Court has acknowledged that there is 'now a tendency to abandon the distinction between jurisdictional facts and other facts which go to the merits', without ever overruling Raza Textiles. The result is a doctrine that says, formally, that jurisdictional facts are reviewable on the High Court's independent judgment, and that says, in operation, that almost any fact can be characterised as jurisdictional if the High Court is inclined to intervene. The distinction has lost its work.

Compare what the post-Anisminic United Kingdom did with the same problem. The House of Lords and Court of Appeal continued, for at least the next two decades, to distinguish between errors that went to jurisdiction and errors within jurisdiction, even as the category of jurisdictional error widened. By the time R v Lord President of the Privy Council, ex parte Page (1993) was decided, the House of Lords had essentially conceded that the distinction had collapsed, and explicitly abandoned it in favour of a general rule that errors of law are reviewable. But it had taken twenty-four years of doctrinal work to reach that point, and the abandonment of the distinction was articulated, defended, and applied as a general rule going forward. In India, the abandonment has been performed, never articulated, and applied selectively. The bench can revive the distinction when it wants to defer, and ignore it when it wants to intervene. That is not a developed doctrine. It is a discretion in a doctrinal costume.

The structural feature: review with no point of finality

TV Somanathan's chapter on the administrative and regulatory state in the Oxford Handbook describes a different, but related, problem: the practical consequence of the doctrine is that there is no point at which a regulatory decision becomes final.¹³ A decision of the Reserve Bank can be reviewed in the High Court. A decision of the Competition Commission appealed to the Competition Appellate Tribunal can be re-challenged in the High Court under Article 226, and then taken to the Supreme Court. A decision of the National Green Tribunal can be challenged in the High Court of the State where the bench sits, and then taken to the Supreme Court — and the Madras High Court has entertained a writ challenge to an order of the Chennai bench of the NGT relating to a regulator in a different State altogether.

The doctrine of review and the doctrine of finality clauses are sides of the same coin. If the second is conceived as a near-absolute floor — a near-presumption that constitutional review survives every legislative attempt at ouster — then the first must be conceived as a structured ceiling, with disciplined categories that determine when review is appropriate and when it is not. India has the floor without the ceiling. The constitutional courts can intervene, in principle, in any administrative decision. The doctrine does not, on its own terms, tell them when they should. Tata Cellular's warning that the court should not 'act as a superboard or with the zeal of a pedantic schoolmaster substituting its judgement for that of the administrator' is recited routinely. It is recited and then disregarded.

Somanathan reads this as a problem of regulatory design: the multiplicity of channels of appeal and the porousness of finality at every level have an economic cost, in increased uncertainty and longer route to a binding outcome. Cases like the Competition Commission's DLF Ltd case have taken from 2010 to a date past 2014 just to reach the Supreme Court on statutory appeal. The economic cost is real. So is the constitutional cost, which is the focus here: an administrative-law system in which the legislature cannot meaningfully signal that a class of decisions should be final, and in which the judiciary cannot meaningfully discipline its own intervention, is a system that does not have the rule-of-law benefits that ouster clauses, properly conceived and properly limited, are designed to confer.

The constitutional shortcut that I gestured at earlier — the basic-structure holding in L Chandra Kumar v Union of India (1997)² — needs a closer look. The constitutional bench held that the power of judicial review under Articles 32 and 226 is part of the basic structure of the Constitution, and that no statute can take it away. The holding has done a great deal of useful constitutional work. It is the holding that limits the reach of tribunalisation and that prevents Parliament from immunising particular subject-matters from constitutional review.

The holding has also, however, done a particular doctrinal damage. By moving the question of ouster clauses out of administrative-law doctrine and into constitutional doctrine, it has relieved the Court of the burden of developing the administrative-law doctrine that ouster clauses, in the Anisminic tradition, were used to develop. The constitutional court does not need to construe the clause carefully to find a route to review. It can simply assert that review is available because review is basic structure. The clause becomes irrelevant.

What the clause might have done, had it been allowed to do work, is what Anisminic made it do. It might have forced the Court to articulate, in each case, the precise jurisdictional ground on which review proceeded. It might have constrained the bench to limit its review to matters that genuinely fall outside the protective scope of the clause, as that scope is properly read. It might have produced, over time, a structured Indian doctrine of jurisdictional error — one that distinguished between substantive ultra vires, procedural ultra vires, jurisdictional-fact error, and error within jurisdiction. India has the vocabulary for all of these. It has the case law gesturing at all of these. What it does not have is a treatment in which the distinctions actually shape the outcome.

Two further consequences follow. First, executive design of regulatory statutes has had to anticipate that any finality clause will be substantially read down. Brahm Dutt v Union of India (2005) and Delhi Science Forum v Union of India (1996) show the executive amending the statute in the course of the case to forestall a basic-structure invalidation.¹² That is regulation-by-anticipated-judicial-veto, and it produces statutes that are written defensively rather than coherently. Second, the legal profession's instinct has shifted in the direction of treating the High Court writ as the first and last forum, even where the legislature has provided an alternative remedy or has named the tribunal as the final adjudicator. The exhaustion-of-alternative-remedies principle survives in the textbook. It is honoured intermittently in practice.

What a disciplined Indian doctrine would look like

I do not want to overstate the prescriptive moment. The constitutional decision in L Chandra Kumar is not going to be revisited, and probably should not be. The basic-structure shortcut is a serviceable constitutional rule. The question is what to do with the administrative-law doctrine that has been allowed to atrophy beneath it.

Three moves seem to me defensible and modest. The first is to articulate the post-Tata Cellular trichotomy as the doctrinal floor for writ review, and to commit to it. Illegality, irrationality, and procedural impropriety should be the categories under which writ review proceeds. 'Arbitrariness' should be retained as a label for the category, not as a freestanding ground that the bench can use to substitute its view of the merits. The Jalan and Rai chapter shows how often the terms have been used interchangeably; the Court could choose to discipline the usage by writing more carefully.

The second is to take the jurisdictional-fact doctrine of Raza Textiles and apply it as a doctrine, not as decoration. If the bench believes that the question before it is a jurisdictional-fact question, it should say so, articulate the basis for treating the fact as jurisdictional, and review the fact on the merits. If it does not, it should leave the matter to the tribunal whose statute conferred the determination on it. The collapse of the distinction has produced a casualness in writ practice that the doctrine does not justify.

The third is to take finality clauses seriously enough that, where the legislature has used clear language to confer finality on a determination, the Court reads the clause carefully before declining to give it effect. L Chandra Kumar does not require the Court to ignore the clause; it requires the Court to acknowledge that constitutional review survives the clause. Acknowledging the survival of constitutional review is consistent with structuring that review to mean something specific. The English doctrine, even after Anisminic, retained respect for the clause as a signal of legislative intent that the tribunal should be the primary adjudicator. The Indian doctrine could do the same without surrendering its constitutional ground.

What is at stake is not the constitutional question of whether the Court can review. It is the administrative-law question of what review consists of, and how the legislature can communicate, through a finality clause, that some classes of decision should be made and stay made without indefinite re-litigation. Anisminic gave the United Kingdom the conceptual tools to answer that question while preserving review. India copied the result and lost the distinction the doctrine was built on. The retrieval of the distinction is the work that an honest Indian doctrine still has to do. The vocabulary is in the cases. The categories are in the textbooks. What is missing is the willingness to commit to them — to say which questions go to jurisdiction and which do not, which kinds of error survive the clause and which do not, and to live with the answers when the next contested administrative determination arrives in writ jurisdiction.

Get the week's biggest decisions, distilled.

A weekly digest for lawyers, students and the legally curious. Every Friday morning. Free.